CN113509191A - Method, device and equipment for analyzing mammary gland molybdenum target X-ray image - Google Patents
Method, device and equipment for analyzing mammary gland molybdenum target X-ray image Download PDFInfo
- Publication number
- CN113509191A CN113509191A CN202110246233.XA CN202110246233A CN113509191A CN 113509191 A CN113509191 A CN 113509191A CN 202110246233 A CN202110246233 A CN 202110246233A CN 113509191 A CN113509191 A CN 113509191A
- Authority
- CN
- China
- Prior art keywords
- image
- diagnosis
- diagnostic
- data
- preset condition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910052750 molybdenum Inorganic materials 0.000 title claims abstract description 44
- 239000011733 molybdenum Substances 0.000 title claims abstract description 44
- ZOKXTWBITQBERF-UHFFFAOYSA-N Molybdenum Chemical compound [Mo] ZOKXTWBITQBERF-UHFFFAOYSA-N 0.000 title claims abstract description 41
- 210000005075 mammary gland Anatomy 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 title claims description 33
- 238000003745 diagnosis Methods 0.000 claims abstract description 113
- 238000004458 analytical method Methods 0.000 claims abstract description 17
- 230000002308 calcification Effects 0.000 claims description 40
- 210000000481 breast Anatomy 0.000 claims description 36
- 210000004907 gland Anatomy 0.000 claims description 24
- 230000008859 change Effects 0.000 claims description 15
- 230000002980 postoperative effect Effects 0.000 claims description 9
- 230000003902 lesion Effects 0.000 claims description 7
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000003384 imaging method Methods 0.000 claims description 6
- 206010006187 Breast cancer Diseases 0.000 abstract description 14
- 208000026310 Breast neoplasm Diseases 0.000 abstract description 14
- 238000012216 screening Methods 0.000 abstract description 11
- 238000001514 detection method Methods 0.000 abstract description 9
- 238000003703 image analysis method Methods 0.000 abstract description 3
- 208000004434 Calcinosis Diseases 0.000 description 39
- 238000007689 inspection Methods 0.000 description 11
- 206010028980 Neoplasm Diseases 0.000 description 10
- 238000010586 diagram Methods 0.000 description 10
- 238000012512 characterization method Methods 0.000 description 9
- 238000011156 evaluation Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 201000011510 cancer Diseases 0.000 description 5
- 210000004027 cell Anatomy 0.000 description 5
- 230000003211 malignant effect Effects 0.000 description 5
- 210000001165 lymph node Anatomy 0.000 description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 230000006866 deterioration Effects 0.000 description 3
- 230000000762 glandular Effects 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 239000008267 milk Substances 0.000 description 3
- 210000004080 milk Anatomy 0.000 description 3
- 235000013336 milk Nutrition 0.000 description 3
- 210000000062 pectoralis major Anatomy 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000004083 survival effect Effects 0.000 description 3
- 230000008719 thickening Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 210000004204 blood vessel Anatomy 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000002405 diagnostic procedure Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 230000009545 invasion Effects 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 210000003205 muscle Anatomy 0.000 description 2
- 210000002445 nipple Anatomy 0.000 description 2
- 238000002601 radiography Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 206010052273 Dystrophic calcification Diseases 0.000 description 1
- 208000031816 Pathologic Dilatation Diseases 0.000 description 1
- 206010040867 Skin hypertrophy Diseases 0.000 description 1
- 210000004883 areola Anatomy 0.000 description 1
- 238000001574 biopsy Methods 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 239000011575 calcium Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005713 exacerbation Effects 0.000 description 1
- 239000007943 implant Substances 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 231100000444 skin lesion Toxicity 0.000 description 1
- 206010040882 skin lesion Diseases 0.000 description 1
- 230000004222 uncontrolled growth Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
- A61B6/5217—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data extracting a diagnostic or physiological parameter from medical diagnostic data
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Optics & Photonics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- High Energy & Nuclear Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Physiology (AREA)
- Pathology (AREA)
- Radiology & Medical Imaging (AREA)
- Physics & Mathematics (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The application provides a mammary gland molybdenum target X-ray image analysis method, which comprises the steps of receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnostic data of the diagnostic categories based on one or more preset diagnostic categories, and sending qualitative diagnostic data to an image structured report according to a third preset condition; the image structured report generates a diagnostic impression based on the entered qualitative diagnostic data. The first image provides an accurate image for projection through screening, results of all diagnosis categories are automatically obtained in subsequent detection, standard diagnosis impressions are output through an image structuralization report, labor is saved, and a standard, reliable and stable breast cancer screening result is provided. The application also provides an analysis device and equipment for the mammary gland molybdenum target X-ray image.
Description
Technical Field
The application relates to an analysis technology of medical images, in particular to an analysis method of mammary gland molybdenum target X-ray images. The application also relates to a mammary gland molybdenum target X-ray image analysis device and equipment.
Background
Breast cancer is one of the most common malignancies in women, caused by uncontrolled growth of breast cells. In general, in the pre-mammary stage, uncontrollable-growing breast cells divide disorderly in the mammary lobules and ducts, eventually forming malignant tumors. Statistically, early breast cancer has a 5-year survival rate of up to 99% after onset, and over time, the 5-year survival rate decreases to 85% once cancer cells have spread to peripheral lymph nodes, and the 5-year survival rate decreases to 27% once cancer cells have metastasized distally. Therefore, early detection and early treatment are the key points for diagnosing and treating breast cancer.
In the prior art, a computerized Radiography-molybdenum target X-ray (CR) is usually adopted to examine breast cancer cells, and the method can relatively comprehensively and accurately reflect the structure of the whole breast in a non-invasive manner and reliably identify whether the breast is a benign lesion or a malignant tumor. As a convenient, low-cost and non-invasive examination method, molybdenum target X-ray has become the first imaging method for breast cancer screening.
In the conventional medical diagnosis, if a patient needs to be diagnosed, registration is firstly needed, then a doctor diagnoses the patient, then the doctor puts out a bill needing to check the body according to the diagnosis result of the doctor, the patient performs a series of body checks according to the bill, finally the doctor determines the body problems of the patient according to the check result, and the determination of the diagnosis result depends on the personal diagnosis capability of the doctor.
In breast cancer molybdenum target X-ray screening, the diagnostic influence of the projection position on the image is serious, and meanwhile, the images in the breast molybdenum target X-ray image are overlapped, so that the difficulty of identifying the image in a line film is increased for a doctor, and the breast density also judges the focus by the image. The number of diagnostic categories for imaging examinations is large, resulting in labor intensive and difficult to quantify. Due to the problems, the current breast cancer screening and diagnosing difficulty is high, and even the misdiagnosis problem is caused at a probability.
Disclosure of Invention
In order to solve the above technical problems, the present application aims to provide an analysis method of breast molybdenum target X-ray images to realize standard, stable and reliable breast cancer diagnosis, and also provides an analysis device and equipment of breast molybdenum target X-ray images.
The application provides an analysis method of mammary gland molybdenum target X-ray images, which comprises the following steps:
receiving a first image;
judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image;
judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnosis data of the diagnosis type based on one or more preset diagnosis types, and sending the diagnosis data to an image structured report according to a third preset condition;
the visual structured report generates a diagnostic impression based on the input diagnostic data.
Optionally, after determining whether the quality parameter of the second image meets a second preset condition, the method further includes: and sending the judgment data to the image structured report.
Optionally, the diagnostic data comprises:
qualitative diagnostic data and quantitative diagnostic data.
Optionally, the qualitative diagnostic data obtaining step is as follows:
calling a recognition model matched with the diagnosis category;
and performing qualitative diagnosis on the second image through the recognition model to generate the qualitative diagnosis data.
Optionally, the identification model includes at least one of an AI model and a cinematology module.
Optionally, the diagnostic categories include:
detecting postoperative change of mammary gland, gland density classification, gland location, calcification, mass, gland symmetry analysis, gland structural distortion, and symbol classification.
Optionally, the method further comprises the following steps:
and comparing the diagnostic data with corresponding historical diagnostic data, and sending the comparison result to the image structured report.
Optionally, after generating the diagnostic data of each diagnostic category, the following steps are performed:
determining a BI-RADS classification of the lesion based on the diagnostic data and the comparison result.
Optionally, after obtaining each of the diagnostic data, the following steps are performed:
storing the diagnostic data in a database of the image structured report.
The application also provides an analysis device for mammary gland molybdenum target X-ray images, which comprises:
a receiving module for receiving a first image;
a first judging module, configured to judge whether a property of the first image meets a first preset condition, and the property of the first image will meet the first preset condition
Defining the first image of a preset condition as a second image;
the second judgment module is used for judging whether the quality parameters of the second image meet second preset conditions or not, if so, generating diagnosis data of the diagnosis categories based on one or more preset diagnosis categories, and sending the diagnosis data to an image structured report according to third preset conditions;
a report analysis module for generating a diagnostic impression from the image structured report based on the input diagnostic data.
The application also provides an analysis device for mammary gland molybdenum target X-ray images, which comprises:
a processor;
a memory storing an image analysis program that, when read by the processor, performs the following: receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnosis data of the diagnosis type based on one or more preset diagnosis types, and sending the diagnosis data to an image structured report according to a third preset condition; the visual structured report generates a diagnostic impression based on the input diagnostic data.
Compared with the prior art, the method has the advantages that:
the application provides a mammary gland molybdenum target X-ray image analysis method, which comprises the steps of receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnostic data of the diagnostic categories based on one or more preset diagnostic categories, and sending qualitative diagnostic data to an image structured report according to a third preset condition; the image structured report generates a diagnostic impression based on the entered qualitative diagnostic data. The first image ensures that the image applied to the examination can provide an accurate image for projection through screening and quality judgment, diagnoses of all diagnosis categories are automatically acquired in subsequent detection items, and a diagnosis impression is output according to a preset standard through an image structured report, so that a standard, reliable and stable breast cancer screening result can be provided while labor is saved.
Drawings
FIG. 1 is a flow chart of the method for analyzing breast molybdenum target X-ray images in the present application.
FIG. 2 is a schematic diagram of an interface for determining image properties in a visual structured report according to the present application.
Fig. 3 is a diagnostic flow chart for each diagnostic category in the present application.
FIG. 4 is a schematic interface diagram of the examination of post-mammary gland changes in an image structured report as described in the present application.
FIG. 5 is a schematic interface diagram of the detection of tumors and characterization of tumors in an image structuring report according to the present application.
Fig. 6 is a schematic interface diagram of detection of calcification and calcification characterization in an image structuring report according to the present application.
FIG. 7 is a flow chart of the call recognition model of the present application.
FIG. 8 is an interface diagram in the present application comparing historical diagnostic results in a structured report of images.
FIG. 9 is a schematic interface diagram of the BI-RADS classification in the image structuring report described in the present application.
FIG. 10 is a schematic view of an apparatus for analyzing X-ray images of breast molybdenum targets according to the present application.
Detailed Description
The following detailed description sets forth specific technical details of the technical solutions of the present application in order to make the objects, features and advantages of the present application more comprehensible. The following description is made for the purpose of illustrating various details of the invention and is not to be taken in a limiting sense. Those skilled in the art can understand the idea of the present invention and make various changes in the implementation means and application scenarios without departing from the protection scope of the present application.
In the embodiment of the application, the images needing to be analyzed are screened and the quality of the images is judged, so that the images can stably and reliably provide focus information, diagnosis data of various diagnosis categories are automatically acquired, and diagnosis impressions are output through image structured reports, so that the labor intensity is saved, and the reliability and the stability of diagnosis images are improved. The method comprises the following steps: receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnosis data of the diagnosis type based on one or more preset diagnosis types, and sending the diagnosis data to an image structured report according to a third preset condition; the visual structured report generates a diagnostic impression based on the input diagnostic data.
Fig. 1 shows a flow chart of the method for analyzing breast molybdenum target X-ray image in the present application, which mainly includes the following steps:
referring to fig. 1, in step S101, a first image is received;
in the present application, the first image may be a molybdenum target X-ray film captured by a Computed Radiography (CR) technique in an imaging department, which has been widely used in the existing practice of image examination, and is mainly used for breast cancer examination. The molybdenum target X-ray can comprehensively and accurately reflect the structure of the whole breast and reliably identify whether the breast is benign lesion or malignant tumor. The first image is mainly a Computed Radiography-molybdenum target X-ray (CR) image for examining the mammary gland.
Step S102, judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image 101;
after receiving a first image, it is necessary to check whether the first image meets a first preset condition. Conventionally, when a molybdenum target X-ray examination is performed on a patient, the patient needs to be registered in advance in an RIS (radiology information management system), where the registered information includes basic information of the patient and examination information, and the examination information is a basis for the molybdenum target X-ray examination performed on the patient.
Fig. 2 shows a schematic interface diagram of image property determination in the visual structured report in the present application.
Referring to fig. 2, in the present application, determining whether the first image meets a first preset condition by using a trained AI model is to determine whether the first image meets inspection information. Wherein the examination information at least comprises information such as the position of the projector, the purpose of examination and the like. For example, as shown in fig. 2, if the breast cancer screening or the diagnostic examination is selected as one of the examination items, the diagnostic examination is schematically selected in this embodiment, and it is determined whether the first image matches with the examination item registered by the RIS when it is determined whether the property of the image matches with the first preset condition.
And judging the first image to obtain a first certainty judgment result, if the first certainty judgment result shows that the first image does not accord with the inspection information, terminating the process, sending second data, recording the second data in a database, and performing subsequent processing by related personnel according to the second data. And when the first certainty judgment result shows that the first image accords with the inspection information, defining the first image as a second image 101, and sending out first data which is returned to a technical evaluation control of the image structured report. In the present application, the second image 101 is an input image for performing a specific diagnosis for each diagnosis category.
Step S103, judging whether the quality parameters of the second image 101 meet a second preset condition, if so, generating diagnosis data of the diagnosis categories based on one or more preset diagnosis categories, and sending qualitative diagnosis data to an image structured report according to a third preset condition;
the first image has been defined as the second image 101 in the above steps according to the first image meeting the first preset condition. In this step, the quality of the second image 101 is mainly checked by the trained recognition model. For example, it is checked whether the picture of the second image 101 is clear, whether the picture has an offset, and the like. In performing an X-ray examination of a molybdenum target, the quality parameters for the image quality are set in the present application to describe the degree of image quality, because the images used for diagnosis of the respective diagnostic categories cannot completely avoid some quality problems for various reasons, such as the inability of the patient to remain absolutely still, positional deviations of the image shots, or for other reasons that cannot be eliminated. And comparing the quality parameter with a preset second preset condition to determine whether the image quality is qualified or not, and sending the judgment data to an image structured report. For example, whether the quality parameter of the second image 101 meets a second preset condition is determined, if yes, the image quality is qualified, and a subsequent process may be performed.
In this application, the second predetermined condition may be a value range, for example, a parameter is selected, so that when the parameter is greater than a certain value, the picture quality is qualified. Or selecting a parameter, so that the picture quality is qualified when the parameter is less than a certain value. And a value interval can be set, so that the picture quality is qualified when the selected parameters are in the value interval. This arrangement may be in other ways, and will not be described in detail here.
And sending third data to a technical evaluation control of the image structured report after the image quality is judged to be qualified, terminating the subsequent flow if the image quality is judged to be unqualified, sending fourth data, recording the fourth data in a database, and performing subsequent processing by related personnel according to the fourth data.
By the above steps, the second image 101 has already finished the evaluation and screening with respect to the image, and in the following steps, a specific diagnostic check will be performed by the second image 101 based on the preset diagnostic category.
Fig. 3 shows a diagnostic flow chart for each diagnostic category in the present application.
Referring to fig. 3, a diagnostic examination is performed based on the second image 101, requiring examination of one or more diagnostic categories. The diagnostic categories are preset and the examination based on the preset diagnostic categories is necessary for each diagnostic examination, but such examination does not exclude incidental findings. In some embodiments, the tests for the diagnosis categories are in a parallel relationship, and in another embodiment, the tests for the diagnosis categories may also be in a sequential relationship, which may be selected by a person skilled in the art according to the actual situation. Preferably, after the detecting the post-mammary gland change outputs diagnostic data, the gland density classification is diagnosed before each of the remaining diagnostic categories. Meanwhile, the diagnosis data of each diagnosis category is sent to an image structured report to realize comprehensive judgment of the focus.
As shown in fig. 3, in the present application, a second image 101 is first entered to examine post-mammary changes 102.
Fig. 4 shows a schematic interface diagram of the examination of post-mammary gland changes in an image structured report as described in the present application.
Referring to fig. 4, after checking the breast post-operation change 102, it can be obtained whether the post-operation change, such as visible post-operation change and invisible post-operation change, and the change position, such as position coordinates, and other diagnostic data. For example, as shown in fig. 4, after examining the breast postoperative change, it can be determined whether the left and right breasts have undergone breast conservation, and whether there are any foreign bodies such as metal clips and prostheses. The mammary gland composition can also be judged, for example, left and right breast classification, and the classification includes: ACR a, ACR b, ACR c, ACR d, and the like. The left and right breasts may be evaluated as a whole, for example, the left breast is not abnormal and the right breast is not abnormal.
In this diagnostic category, the content of the main examinations includes: the post-operative altered area is segmented while also identifying foreign objects such as metal clips, prosthetic implants, etc.
In the present application, a trained recognition model is used to check for breast postoperative changes, and postoperative region coordinates and foreign object coordinates, such as coordinates of a metal clip tag, are obtained from the second image 101. And finally, extracting a key image of the postoperative change area according to a preset rule, and returning a qualitative judgment result and the key image to a related technology control of the image structured report. It should be noted that if the qualitative determination result is that no change is found after the operation, the qualitative determination result may not be returned to the structured report, and those skilled in the art may perform reasonable setting according to the specific actual situation.
By examining the post-operative changes 102 of the breast, the post-operative change area can be accurately found, on the basis of which further diagnosis can be made.
The diagnostic categories of the further diagnosis include: gland density classification 103, gland localization 105, examination calcification 109, calcification characterization 1091, examination mass 108, mass characterization 1081, gland symmetry analysis 107, examination of gland structural distortion 104, and symbol classification 106. In the diagnosis of each diagnosis category, the second image 101 may be diagnosed by using different recognition models according to actual needs. In the present application, the identification model at least includes one of an AI model and a proteomics module, and each diagnosis category corresponds to a dedicated AI model and/or a proteomics module. And interfacing the diagnosis data with the system through a program or a script so that the obtained result automatically enters an inspection process, and forming a final diagnosis impression through the diagnosis data by means of an image structured report.
For the detection categories, the gland density classification 103 is firstly carried out, and the diagnosis of other categories can be influenced by high-density mammary glands, so that after the density classification is determined, whether the mammary gland density influences the diagnosis of the focus can be judged according to specific conditions, and the best diagnosis effect can be obtained. In the present application, the gland density classification 103 includes: lipidated, fibroglandular, non-uniform glandular, and compact.
In the present application, diagnosis of calcification 1091 or mass 1081 requires examination of calcification 109 or mass 108, respectively, and only examination of calcification 1091 or mass 1081 is considered.
FIG. 5 shows a schematic interface of the detected masses 108 and qualitative masses 1081 in the image structured report.
Referring to fig. 5, when the detected tumor 108 is determined to have a tumor in the milk, a tumor characterization 1081 is performed, wherein the tumor characterization includes: number, shape, size, edges, and density. The number may be expressed as: single or multiple; the shape may be expressed as: circular, oval, or irregular; the size may be represented in a multi-parameter representation, for example, using a first parameter radial line and a second parameter maximum radial line or range; the edge may be represented as: sharp, shadowing, fuzzy, differential paging or burring; the density can be expressed as: high density, equal density, low density, or fat containing density, etc.
Fig. 6 shows a schematic interface of the detection of calcification 109 and calcification characterization 1091 in the image structuring report of the present application.
Referring to FIG. 6, when calcification 109 is detected and it is determined that there is calcification in milk 1091, the calcification determination includes: morphology, distribution and calcification range. The morphology shown can be expressed as: typically benign calcification or suspicious malignant calcification. The typical benign calcification or suspicious malignant calcification shown has one or more sub-options, which may be single-choice or multiple-choice, with specific choices depending on the specific diagnosis. Sub-options for the typical benign calcification include: skin calcification, which can further judge the calcification types of the skin calcification, and is respectively thick or popcorn-like calcification, round calcification, edge type calcification and calcium milk calcification; and (3) calcification of the blood vessel, wherein the calcification types of the blood vessel can be further judged and are large rod-shaped calcification, punctate calcification, dystrophic calcification and suture calcification respectively. The distribution can be expressed as: diffuse, regional, cluster, line-like, segment-like. The calcification ranges may be represented by specific numerical range edges.
A symbolic classification 106, which is an additional or adjunct diagnostic category found while other diagnostic categories were in progress. Based on other problems found by the normal diagnostic procedure, the other problems need to be classified, for example, the accompanying metaphor may be: thickening of the skin, skin retraction, nipple retraction, enlargement of axillary lymph nodes, thickening of the trabeculae or invasion of the pectoralis major muscle, etc., either visible or not in a particular examination; other symbolic classifications may include: lymph nodes within the breast, skin lesions, isolated ductal ectasia, etc., are shown visible or not visible in the detailed examination.
In the present application, after the above diagnosis order is met, the execution order of other diagnosis categories may be performed without being sequenced.
The examination order of each examination item described in detail above can be used to acquire diagnostic data for each diagnosis type by performing an examination based on the examination order. In the present application, the diagnostic data includes qualitative diagnostic data and quantitative diagnostic data, the qualitative diagnostic data is obtained by comparing the recognition result of the second image with the comparison result through the recognition model, and the comparison result may be a parameter indicating that the examination is normal, or a historical diagnostic image, etc. The quantitative diagnostic data is measurement result data obtained by measuring each distinguished region in the image based on the identification of the image.
In the present application, each screening diagnosis of breast cancer may involve a plurality of diagnostic categories, but the numerical properties of these diagnostic categories that need to be measured may differ, for example, measurements including density, length, or area may be measured. Therefore, different recognition models are provided for accurately diagnosing different diagnostic categories.
FIG. 7 is a flow chart of the call recognition model of the present application.
Referring to fig. 7, the following steps are performed when performing diagnosis of a specific diagnosis category:
s201, calling an identification model corresponding to the diagnosis category;
s202 performs qualitative diagnosis on the second image 101 by using the recognition model, and generates diagnosis data.
After the above diagnosis is completed and the diagnosis results of each diagnosis category are obtained, further analysis is performed and the diagnosis results are compared with historical diagnosis results.
Fig. 8 shows an interface diagram in the image structured report compared with the historical diagnosis result in the present application.
As shown in fig. 8, the diagnostic data is compared with the corresponding historical diagnostic data and the comparison result is sent to the image structured report. The comparison result includes the following information: inspection time, image source, inspection method, comparison result. In the method, firstly, a diagnosis result of each diagnosis category is obtained according to an identification model, and whether the diagnosis result is a positive result is judged; and then, if the diagnosis result is positive, determining the diagnosis data according to the positive result and the corresponding historical diagnosis result. The quantitative diagnostic data in the present application is measurement result data, that is, data generated by measuring an identification model for each diagnostic category.
Wherein, the image source can be home, outer hospital and others; the examination methods shown include breast molybdenum targets and others; the comparison results include no obvious change and previous change, and if the previous change is changed, the text description can be carried out.
After generating the diagnostic data, each diagnostic category of the present application will send the diagnostic data to the image structured report according to a third preset condition. The third preset condition refers to a sending condition of the diagnosis data, and includes position coordinates and the like of a sending purpose of the diagnosis data in the image structured report.
In the specific diagnosis step of the diagnosis category, the second image 101 is input first, the identification model, such as an AI model and/or a cinematology module, is used to perform the diagnosis of the lesion, and the diagnosis data of each diagnosis category is obtained.
To further describe the details of the qualitative judgments, the present application classifies the qualitative judgments for the diagnostic categories into 4 categories, the qualitative judgments belonging to at least the following categories: image feature variation, image feature density, image feature positioning and image feature classification.
The image feature variation degree is a comparison result obtained by comparing a currently measured image feature with a historical image feature, or a comparison result of a conventionally checked result of the currently measured image feature and the image feature. The specific checking mode is as follows: and acquiring image characteristic attribute features such as density, coordinates and the like. And comparing the acquired attribute features with historical attribute features to acquire a comparison result. Preferably, the historical attribute feature is a previous attribute feature.
The image feature density refers to the number or density of the currently measured image features, such as the density of the gland in the mammary gland. The specific checking mode is as follows: and acquiring the position coordinates of the image features, determining the image feature range according to the position coordinates, and measuring the density of the image features in the image feature range.
The image feature positioning refers to coordinates of the currently measured image features in an entity shown in the image. The specific checking mode is as follows: and acquiring the image characteristic position coordinates through the recognition model.
The image feature classification refers to classification or degree of currently measured image features at a technical level. The specific checking mode is as follows: and measuring the image feature description data through a recognition model, and determining the category of the image feature according to the description data.
The above classification is made in the present application for the purpose of classifying the diagnostic methods of the respective diagnostic categories and the characteristics of the acquisition of the diagnostic data, and is not specifically embodied in the practical application, and does not limit the scope of protection of the present application. Meanwhile, the 4 categories may be related to each other or crossed, or may be independent.
The qualitative determination of each diagnostic category is explained below:
the gland density classification 104 belongs to image feature density and image feature classification, the inspection result is corresponding gland density data, and the classification category can be determined according to the density data. The classified categories may be represented as: lipidated, fibroglandular, non-uniform glandular, dense.
The gland positioning 106 belongs to image feature positioning, the inspection result is coordinate data, the image position can be determined according to the coordinate data, and the position can be represented as the right side and the left side; inner quadrant, outer quadrant, upper quadrant, lower quadrant; front, middle, rear; behind the areola; the axillary caudal region; pectoralis major, etc.
The examination calcification 108 belongs to image feature positioning and image feature variation, and the examination result is as follows: coordinate data and whether calcification has occurred.
The calcification characterization 1081 belongs to an image feature classification, which may be typical benign calcification, suspicious malignant calcification, and the shape, distribution characteristics, etc. of calcification.
The detected tumor 105 belongs to image feature positioning and image feature variation, and the detection result is as follows: coordinate data and whether a lump occurred.
The mass characterization 1051 belongs to the image feature classification, and the classification of the mass may be: typical benign masses, suspicious malignant masses, and the morphology, distribution characteristics of the masses.
The gland symmetry classification 107 belongs to image feature classification, and the inspection result may be: symmetrical and asymmetrical.
The detected glandular structure distortion 104 belongs to the image feature classification, and the detection result can be: the structural distortion is visible and not visible.
The accompanying symbol classification and other symbol classification 106 can perform qualitative classification of the above 4 categories based on the characteristics of the actual image features, including skin thickening, skin retraction, nipple retraction, axillary lymph node enlargement, trabecular thickening, pectoralis major muscle invasion, and the examination result can be described as: visible and invisible.
The above steps have acquired a diagnosis for each diagnostic category, and based on the qualitative diagnostic data, BI-RADS (breast image report and data system) classification of the lesion is determined. For example, the BI-RADS classification of the lesion is determined from the coordinates of the metal clip tag, the localization code of the gland location, the coordinates of the calcification tag, the coordinates of the mass tag, the coordinates of the structural distortion tag.
FIG. 9 shows a schematic interface diagram of the BI-RADS classification described in the present application in an image structuring report.
Referring to FIG. 9, the BI-RADS classification includes complete evaluation and incomplete evaluation. The complete assessment includes six categories, respectively: class 1, negative, almost zero probability of deterioration; class 2, positive, almost zero; class 3, probability of deterioration > 0% but < 2%; class 4, deterioration probability > 2% but ≦ 95%; class 5, exacerbation probability > 95%; class 6, biopsy results are known to be malignant. The categories of incomplete evaluation include: class 0, which requires further imaging evaluation and/or comparison of previous examination results.
And step S104, generating a diagnosis impression according to the input qualitative diagnosis data by the image structured report.
In the above steps, the diagnostic data of each diagnostic category is acquired by diagnosis of each diagnostic category. In the step, in the image structured report, the final diagnosis impression is obtained by combining all the data, and the diagnosis data generated in the diagnosis process is stored in an image structured report database.
The above embodiments fully describe the method for analyzing breast molybdenum target X-ray images in the present application, and first, screening the image to be examined, and examining various diagnostic categories based on the qualified images. The application also provides an image analysis device corresponding to the image analysis method.
Fig. 10 is a schematic view of an analysis apparatus for breast molybdenum target X-ray images in the present application, which mainly includes a receiving module, a first determining module, a second determining module, and a report analyzing module.
Referring to fig. 10, a receiving module 201 is configured to receive a first image;
in the present application, the first image may be a molybdenum target X-ray film taken by an imaging department through a computerized radiography technique, which is mainly applied to the examination of breast cancer. The first image is mainly a Computed Radiography-molybdenum target X-ray (CR) image for examining the mammary gland.
A first determining module 202, configured to determine whether a property of the first image meets a first preset condition, and define the first image meeting the first preset condition as a second image 101;
after receiving the first image, the first determining module needs to check whether the first image meets a first preset condition. Conventionally, when a molybdenum target X-ray examination is performed on a patient, the patient needs to be registered in the RIS in advance, wherein the registered information includes basic information and examination information of the patient, and the examination information is a basis for the molybdenum target X-ray examination performed on the patient. In the application, whether the first image meets the first preset condition or not is judged by using the trained AI model, namely whether the first image meets the check information or not is judged. Wherein the examination information at least comprises information such as the position of the projector, the purpose of examination and the like. And judging the first image to obtain a first certainty judgment result, if the first certainty judgment result shows that the first image does not accord with the inspection information, terminating the process, sending second data, recording the second data in a database, and performing subsequent processing by related personnel according to the second data. And when the first certainty judgment result shows that the first image accords with the inspection information, defining the first image as a second image 101, and sending out first data which is returned to a technical evaluation control of the image structured report.
A second determining module 203, configured to determine whether the quality parameter of the second image 101 meets a second preset condition, if yes, generate diagnostic data of the diagnostic category based on one or more preset diagnostic categories, and send qualitative diagnostic data to the image structured report according to a third preset condition;
the second determination module mainly checks the quality of the second image 101 through a trained recognition model. For example, it is checked whether the picture of the second image 101 is clear, whether the picture has an offset, and the like. In the molybdenum target X-ray examination, images for diagnosis of each diagnosis category cannot completely avoid some quality problems due to various reasons that cannot be eliminated, so quality parameters are set for image quality in the present application for describing the degree of picture quality. And comparing the quality parameter with a preset second preset condition to determine whether the image quality is qualified or not, and sending the judgment data to an image structured report. For example, whether the quality parameter of the second image 101 meets a second preset condition is determined, if yes, the image quality is qualified, and a subsequent process may be performed.
The second judging module further comprises:
the diagnosis module 2031, after determining that the quality of the image is acceptable, performs one or more diagnostic type checks based on the second image 101 and sends the diagnostic data to the image structured report of the report analysis module.
A report analysis module 204 for generating a diagnostic impression from the image structured report based on the inputted qualitative diagnostic data.
And combining all the diagnosis data to obtain a final diagnosis impression in the image structured report, and storing the diagnosis data generated in the diagnosis process into an image structured report database.
The application also provides an analytical equipment of mammary gland molybdenum target X ray piece, its characterized in that includes: a processor; a memory storing an image analysis program that, when read by the processor, performs the following: receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnostic data of the diagnostic categories based on one or more preset diagnostic categories, and sending qualitative diagnostic data to an image structured report according to a third preset condition; the image structured report generates a diagnostic impression based on the entered qualitative diagnostic data.
Claims (11)
1. A method for analyzing mammary gland molybdenum target X-ray images is characterized by comprising the following steps:
receiving a first image;
judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image;
judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnosis data of the diagnosis type based on one or more preset diagnosis types, and sending the diagnosis data to an image structured report according to a third preset condition;
the visual structured report generates a diagnostic impression based on the input diagnostic data.
2. The method for analyzing the mammary molybdenum target X-ray image according to claim 1, wherein after determining whether the quality parameter of the second image meets a second predetermined condition, the method further comprises: and sending the judgment data to the image structured report.
3. The method for analyzing breast molybdenum target X-ray images of claim 1, wherein said diagnostic data comprises:
qualitative diagnostic data and quantitative diagnostic data.
4. The method for analyzing breast molybdenum target X-ray images according to claim 3, wherein the qualitative diagnostic data is obtained by the steps of:
calling a recognition model matched with the diagnosis category;
and performing qualitative diagnosis on the second image through the recognition model to generate the qualitative diagnosis data.
5. The method of claim 4, wherein the identification model comprises at least one of an AI model and an imaging omics module.
6. The method for analyzing breast molybdenum target X-ray images of claim 1, wherein said diagnostic categories comprise:
detecting postoperative change of mammary gland, gland density classification, gland location, calcification, mass, gland symmetry analysis, gland structural distortion, and symbol classification.
7. The method for analyzing breast molybdenum target X-ray images of claim 1, further comprising the steps of:
and comparing the diagnostic data with corresponding historical diagnostic data, and sending the comparison result to the image structured report.
8. The method for analyzing breast molybdenum target X-ray images according to claim 7, wherein the following steps are performed after the diagnostic data for each diagnostic category is generated:
determining a BI-RADS classification of the lesion based on the diagnostic data and the comparison result.
9. The method for analyzing breast molybdenum target X-ray images according to claims 1-8, wherein the following steps are performed after each diagnostic data is obtained:
storing the diagnostic data in a database of the image structured report.
10. An apparatus for analyzing breast molybdenum target X-ray images, comprising:
a receiving module for receiving a first image;
the first judging module is used for judging whether the property of the first image meets a first preset condition or not and defining the first image meeting the first preset condition as a second image;
the second judgment module is used for judging whether the quality parameters of the second image meet second preset conditions or not, if so, generating diagnosis data of the diagnosis categories based on one or more preset diagnosis categories, and sending the diagnosis data to an image structured report according to third preset conditions;
a report analysis module for generating a diagnostic impression from the image structured report based on the input diagnostic data.
11. An apparatus for analyzing X-ray images of breast molybdenum targets, comprising:
a processor;
a memory storing an image analysis program that, when read by the processor, performs the following: receiving a first image; judging whether the property of the first image meets a first preset condition or not, and defining the first image meeting the first preset condition as a second image; judging whether the quality parameters of the second image meet a second preset condition, if so, generating diagnosis data of the diagnosis type based on one or more preset diagnosis types, and sending the diagnosis data to an image structured report according to a third preset condition; the visual structured report generates a diagnostic impression based on the input diagnostic data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110246233.XA CN113509191A (en) | 2021-03-05 | 2021-03-05 | Method, device and equipment for analyzing mammary gland molybdenum target X-ray image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110246233.XA CN113509191A (en) | 2021-03-05 | 2021-03-05 | Method, device and equipment for analyzing mammary gland molybdenum target X-ray image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113509191A true CN113509191A (en) | 2021-10-19 |
Family
ID=78061137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110246233.XA Pending CN113509191A (en) | 2021-03-05 | 2021-03-05 | Method, device and equipment for analyzing mammary gland molybdenum target X-ray image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113509191A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114710657A (en) * | 2022-04-07 | 2022-07-05 | 北京医准智能科技有限公司 | Evaluation system and method for mammary X-ray image quality |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN108464840A (en) * | 2017-12-26 | 2018-08-31 | 安徽科大讯飞医疗信息技术有限公司 | A kind of breast lump automatic testing method and system |
| CN109285147A (en) * | 2018-08-30 | 2019-01-29 | 北京深睿博联科技有限责任公司 | Image processing method and device, server for breast molybdenum target calcification detection |
| CN111161822A (en) * | 2018-11-08 | 2020-05-15 | 北京赛迈特锐医疗科技有限公司 | Medical image artificial intelligence diagnosis data management system and management method thereof |
| CN111709950A (en) * | 2020-08-20 | 2020-09-25 | 成都金盘电子科大多媒体技术有限公司 | Mammary gland molybdenum target AI auxiliary screening method |
| CN112263236A (en) * | 2020-09-22 | 2021-01-26 | 北京赛迈特锐医疗科技有限公司 | Whole-body tumor MRI intelligent evaluation system and method |
-
2021
- 2021-03-05 CN CN202110246233.XA patent/CN113509191A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108464840A (en) * | 2017-12-26 | 2018-08-31 | 安徽科大讯飞医疗信息技术有限公司 | A kind of breast lump automatic testing method and system |
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN109285147A (en) * | 2018-08-30 | 2019-01-29 | 北京深睿博联科技有限责任公司 | Image processing method and device, server for breast molybdenum target calcification detection |
| CN111161822A (en) * | 2018-11-08 | 2020-05-15 | 北京赛迈特锐医疗科技有限公司 | Medical image artificial intelligence diagnosis data management system and management method thereof |
| CN111709950A (en) * | 2020-08-20 | 2020-09-25 | 成都金盘电子科大多媒体技术有限公司 | Mammary gland molybdenum target AI auxiliary screening method |
| CN112263236A (en) * | 2020-09-22 | 2021-01-26 | 北京赛迈特锐医疗科技有限公司 | Whole-body tumor MRI intelligent evaluation system and method |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114710657A (en) * | 2022-04-07 | 2022-07-05 | 北京医准智能科技有限公司 | Evaluation system and method for mammary X-ray image quality |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10984528B2 (en) | System and apparatus for clinical decision optimisation | |
| Lakhman et al. | Differentiation of uterine leiomyosarcoma from atypical leiomyoma: diagnostic accuracy of qualitative MR imaging features and feasibility of texture analysis | |
| Al Mohammad et al. | A review of lung cancer screening and the role of computer-aided detection | |
| Han et al. | Volume versus diameter assessment of small pulmonary nodules in CT lung cancer screening | |
| CN108898595B (en) | Construction method and application of positioning model of focus region in chest image | |
| Chakraborty et al. | Observer studies involving detection and localization: modeling, analysis, and validation | |
| Youk et al. | Automated volumetric breast density measurements in the era of the BI-RADS fifth edition: a comparison with visual assessment | |
| US9047664B2 (en) | Apparatus and method for utilizing mammogram images for verification | |
| US9177379B1 (en) | Method and system for identifying anomalies in medical images | |
| US20050201599A1 (en) | Diagnostic imaging support apparatus and diagnostic imaging support method | |
| JP2004000609A (en) | Computer assisted diagnosis by multiple energy image | |
| JPH06121792A (en) | Method for analyzing misdiagnosis on ordinary example in automatic detection of lung micronodulation in digital thoracic x-ray photographand its system | |
| Packard et al. | Effect of slice thickness on detectability in breast CT using a prewhitened matched filter and simulated mass lesions | |
| CN101626726A (en) | Identification and analysis of lesions in medical imaging | |
| López et al. | Digital image analysis in breast cancer: an example of an automated methodology and the effects of image compression | |
| WO2012082994A2 (en) | Method for assessing breast density | |
| KR20150073628A (en) | System and method for adapting diagnosis model of computer aided diagnosis | |
| US20210217166A1 (en) | Automated screening of medical data | |
| US20070223807A1 (en) | Medical imaging visibility index system and method for cancer lesions | |
| US10776666B2 (en) | Apparatus and method for diagnosis of medical image | |
| Sawada et al. | Digital volumetric measurement of mammographic density and the risk of overlooking cancer in Japanese women | |
| CN113509191A (en) | Method, device and equipment for analyzing mammary gland molybdenum target X-ray image | |
| CN113100742B (en) | Mammary gland MR image intelligent diagnosis method, device and equipment | |
| WO2008111071A1 (en) | Modifying software to cope with changing machinery | |
| CN111265234A (en) | Method and system for judging properties of lung mediastinal lymph nodes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |