CN111899880A - Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method - Google Patents
Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method Download PDFInfo
- Publication number
- CN111899880A CN111899880A CN202010769350.XA CN202010769350A CN111899880A CN 111899880 A CN111899880 A CN 111899880A CN 202010769350 A CN202010769350 A CN 202010769350A CN 111899880 A CN111899880 A CN 111899880A
- Authority
- CN
- China
- Prior art keywords
- lumbar
- trabecular bone
- trabecular
- bone
- image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000008859 change Effects 0.000 title claims abstract description 39
- 238000012502 risk assessment Methods 0.000 title claims abstract description 14
- 210000000988 bone and bone Anatomy 0.000 claims abstract description 139
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 21
- 230000036541 health Effects 0.000 claims abstract description 17
- 230000000877 morphologic effect Effects 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000012216 screening Methods 0.000 claims abstract description 10
- 206010061363 Skeletal injury Diseases 0.000 claims abstract description 6
- 208000010392 Bone Fractures Diseases 0.000 claims description 28
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 15
- 201000010099 disease Diseases 0.000 claims description 14
- 238000004458 analytical method Methods 0.000 claims description 11
- 238000003709 image segmentation Methods 0.000 claims description 5
- 238000006073 displacement reaction Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 239000011664 nicotinic acid Substances 0.000 claims description 4
- 231100000915 pathological change Toxicity 0.000 claims description 4
- 230000036285 pathological change Effects 0.000 claims description 4
- 206010061728 Bone lesion Diseases 0.000 claims description 3
- 208000032170 Congenital Abnormalities Diseases 0.000 claims description 3
- 206010061619 Deformity Diseases 0.000 claims description 3
- 206010028980 Neoplasm Diseases 0.000 claims description 3
- 239000002639 bone cement Substances 0.000 claims description 3
- 230000007717 exclusion Effects 0.000 claims description 3
- 208000027866 inflammatory disease Diseases 0.000 claims description 3
- 230000003902 lesion Effects 0.000 claims description 3
- 210000004705 lumbosacral region Anatomy 0.000 claims description 3
- 206010039722 scoliosis Diseases 0.000 claims description 3
- 238000001356 surgical procedure Methods 0.000 claims description 3
- 208000003044 Closed Fractures Diseases 0.000 claims description 2
- 230000004075 alteration Effects 0.000 claims 3
- 230000008602 contraction Effects 0.000 claims 1
- 238000003745 diagnosis Methods 0.000 abstract description 15
- 206010017076 Fracture Diseases 0.000 description 25
- 238000011160 research Methods 0.000 description 13
- 230000008569 process Effects 0.000 description 7
- 230000006378 damage Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 238000013425 morphometry Methods 0.000 description 5
- 208000027418 Wounds and injury Diseases 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 208000014674 injury Diseases 0.000 description 4
- 238000010603 microCT Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 210000001519 tissue Anatomy 0.000 description 4
- 206010069689 Spinal column injury Diseases 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 210000000629 knee joint Anatomy 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 206010051763 Bone marrow oedema Diseases 0.000 description 1
- 206010010214 Compression fracture Diseases 0.000 description 1
- 206010060820 Joint injury Diseases 0.000 description 1
- 208000008930 Low Back Pain Diseases 0.000 description 1
- 208000013201 Stress fracture Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000013529 biological neural network Methods 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000006806 disease prevention Effects 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000002595 magnetic resonance imaging Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000004660 morphological change Effects 0.000 description 1
- 230000003562 morphometric effect Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000000399 orthopedic effect Effects 0.000 description 1
- 201000008482 osteoarthritis Diseases 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30008—Bone
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention relates to a lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method, which comprises the steps of collecting lumbar X-ray images of a lumbar patient, screening and carrying out image processing, and establishing a lumbar vertebral body identification model; acquiring a trabecular bone image from the image of the lumbar vertebra vertebral body recognition model, analyzing morphological parameters and change rules of the trabecular bone, and measuring and evaluating the structural health condition of the trabecular bone; based on the health condition of the trabecular bone structure, images of healthy trabeculae, hidden fractures and trabecular bone injuries in the trabecular bone image are identified, learning and training are carried out on the convolutional neural network through different types of trabecular bone images, and the category of the trabecular bone image is automatically identified through the trained convolutional neural network. By the method, the load of the trabecular bone and the fracture can be judged more accurately, meanwhile, more abundant types of image data can be judged and processed in a large batch, the basic labor of doctors can be greatly released, the diagnosis efficiency is improved, and the medical cost is saved.
Description
Technical Field
The invention relates to the technical field of machine learning, in particular to an artificial risk assessment method for lumbar vertebra trabecular load stress change and hidden fracture.
Background
Military training is an important means to transform ordinary people into soldiers with considerable combat ability. In recent years, non-combat derepreneur events caused by excessive training or unscientific training of military forces of countries in the world are frequent. Investigation researches show that officers and soldiers in troops have a large proportion of lumbar diseases, and particularly field officers and soldiers are more serious. The military training with overload is easy to cause training injuries such as stress fracture, knee joint injury, training lower back pain and the like of soldiers. Military training should enhance medical monitoring and objectively measure the training amount through scientific data.
In a modern battlefield, the success of a war depends largely on whether military soldiers have good weaponry. Generally speaking, the more advanced the arming of soldier equipment is, the more likely it is to gain the initiative of the battlefield, but this does not mean that the more weaponry is, the more advantageous it is, because everyone has a limited load capacity. Although soldiers in a battlefield are trained and may have several times or even ten times higher loading capacity than ordinary people, this does not represent that the loading capacity of soldiers is infinite. This is as if we were playing a game, although it is possible to increase the backpack capacity of oneself, the range of capacity increase is always upper-bound. If the upper limit is exceeded, the weapon can not be prepared in the game, and the overload load is often caused to the soldiers to deform and the like on the real battlefield.
Taking the weight of American military soldier equipment as an example, in an environment of real-time combat, the weight of a load of a general soldier is 28.6kg, and the weight of the load is 45.7kg under a marching state, but the weight of the load is increased to 59.8kg under the condition of lack of supply. Due to the limitation of human physiological structures, the load bearing capacity of soldiers has an upper limit, and in order to enhance the load bearing capacity, mechanical exoskeletons are even developed in the United states to help soldiers increase the load bearing capacity of the soldiers, but at present, the research result is not mature and is not popularized and applied. Despite the presence of mechanical exoskeletons and the intensive support of military training, the U.S. military is still faced with the problem of skeletal injury disease caused by excessive loading of soldiers. The results of existing research show that the number of soldiers in the U.S. army retired due to musculoskeletal-related diseases caused by excessive weight bearing has increased ten-fold over the six years since 2003, and the U.S. expenses paid to treat such soldiers for related diseases have reached $ 5 million each year.
As early as 90 years of the last century, the world health organization predicted that "medicine in the 21 st century, should not continue to be the subject of major research on diseases, but rather human health, as the leading direction of medical research". The trend of medical development has been shifted from "unlimited pursuit of high technology for the purpose of treatment", to "prevention of diseases and injuries, maintenance and improvement of health level". In modern society, more energy should be put into the work of keeping healthy and preventing diseases, and unnecessary waste of medical resources is reduced through the guarantee of personal health. In short, it is "disease-free and disease-rarely-caused".
The military training intensity and frequency of the field troops are the greatest by combining the military national conditions of China, most of military institutions are located in remote areas, the military institutions are often inconvenient to traffic, medical and health equipment and peripheral medical resources cannot catch up with the grade-III hospitals in the prefecture, many basic health fleets are basically only equipped with common X-ray equipment, and the field square cabins and the field medical teams are also only equipped with mobile X-ray equipment. The X-ray flat sheet can be diagnosed only when obvious lumbar compression fracture or lumbar isthmus fissure occurs, and hidden trabecular bone fracture is not found positively, so that the condition of the patient is delayed, and the best diagnosis and treatment time is lost. At present, for the image diagnosis of the trabecular hidden fracture of the bone joint, the most sensitive Imaging device is Magnetic Resonance Imaging (MRI), and the device can timely find the bone marrow edema of the trabecular hidden fracture under the condition that no morphological change is found in the trabecular hidden fracture. However, the examination of lumbar vertebrae and knee joints of the equipment is long in time consumption and expensive, and is mostly concentrated in a third-class hospital in the prefecture level, for field troops in remote areas, military training personnel are more involved, tasks are heavy, injuries and diseases are trained, the examination of hospitalizing and examining for leave-on needs to be checked in multiple levels, and in addition, the traffic is inconvenient, injured officers and soldiers often cannot be effectively diagnosed and treated in the first time, so that unnecessary non-combat personnel reduction is caused to the troops, and the individuals of the officers and soldiers involved in the training are injured and even regretted for the whole life.
Disclosure of Invention
The invention aims to solve the technical problem of providing an artificial risk assessment method for lumbar vertebra trabecular load stress change and hidden fracture aiming at the defects of the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows: a lumbar vertebra trabecular load stress change and concealed fracture artificial risk assessment method is constructed, and comprises the following steps:
collecting lumbar X-ray images of a lumbar patient, screening and processing the images, and establishing a lumbar vertebral body identification model;
acquiring a trabecular bone image from the image of the lumbar vertebra vertebral body recognition model, analyzing morphological parameters and change rules of the trabecular bone, and measuring and evaluating the structural health condition of the trabecular bone;
based on the health condition of the trabecular bone structure, images of healthy trabeculae, hidden fractures and trabecular bone injuries in the trabecular bone image are identified, learning and training are carried out on the convolutional neural network through different types of trabecular bone images, and the category of the trabecular bone image is automatically identified through the trained convolutional neural network.
Wherein, in the step of screening the lumbar vertebrae X-ray image of the lumbar vertebrae patient, the inclusion standard includes:
the lumbar X-ray image shooting time point is within three months from the current time point, and the corresponding patient does not receive treatment affecting BMD;
the lumbar X-ray image includes a front view and a side view, including at least images of the first to fourth lumbar vertebrae.
Wherein, in the step of screening lumbar vertebrae X-ray image to lumbar vertebrae patient, the exclusion standard includes:
the patient corresponding to the lumbar X-ray image is subjected to lumbar surgery with internal fixation or bone cement filling;
at least one of the first to fourth lumbar vertebrae represents a lesion of a tumor, inflammatory disease, or scoliosis or deformity;
the lumbar X-ray image cannot be matched with the mask;
the resolution of the lumbar X-ray image is low.
Wherein, the morphological parameters of the trabecular bone at least comprise:
bone volume, characterizing trabecular bone volume within the target region;
a total volume, representing the total volume of the target region;
volume fraction, representing the ratio of trabecular bone volume to total volume;
bone surface area, characterizing the surface area of trabeculae of bone in the target region;
specific surface area, which characterizes the total area per unit volume;
trabecular bone thickness, which characterizes the average thickness between beam structures;
trabecular bone gap, representing the average distance between beam structures, when increased, the display distance is increased, the structure is reduced, and the performance is reduced;
the number of trabeculae represents the number of intersection points of the beam units and the non-beam structures, and is related to thickness and gap variation;
the structural model index represents the composition conditions of trabecular bone and columnar beam, when bone pathological changes occur, the trabecular bone is reduced, the columnar trabecular bone is increased, and the structural model index is increased;
connectivity, characterizing the interconnection condition of the beam structure;
anisotropy represents the ratio of the long axis to the short axis in the intercept length fitting ellipsoid in the target region, and the anisotropy value of the load-bearing trabecula bone is increased at the initial stage of bone lesion and is reduced along with the aggravation of the disease.
Wherein, the trabecular bone is compressed along the axial direction, the structural morphological parameter analysis and the bionic reconstruction shrink or stretch of the trabecular bone, and the structural morphology of the trabecular bone is consistent with the quantitative distribution of the mechanical load.
Wherein, after the step of training the convolutional neural network, the method further comprises the step of evaluating the network energy efficiency by using an observer operating characteristic curve (ROC curve). The accuracy is reflected by the size of AUC of the area under the ROC curve, and the AUC is less than or equal to 0.8, so that the accuracy is low; 0.8< AUC ≦ 0.9 indicating moderate accuracy; AUC >0.9, indicating high accuracy.
Wherein, the two-dimensional graph of the trabecular bone is processed by a convolutional neural network algorithm to identify displacement, scaling and other form distortion invariance.
Wherein, in the step of performing image processing on the lumbar X-ray image, a trabecular bone image in the lumbar X-ray image is acquired by an image segmentation method.
Furthermore, the invention constructs a computer device, comprising an input/output unit, a memory and a processor, wherein the memory stores computer readable instructions, and the computer readable instructions, when executed by the processor, cause the processor to execute the steps of the lumbar vertebra trabecular load stress change and concealed fracture artificial risk assessment method according to the technical scheme.
Furthermore, the present invention constructs a storage medium storing computer readable instructions, which when executed by one or more processors, cause the one or more processors to perform the steps of the method for artificial risk assessment of load stress change of lumbar vertebrae trabecular and concealed fracture as described in the above technical solution.
According to the artificial risk assessment method for load stress change of the trabecular bone of the lumbar vertebra and hidden fracture, a lumbar vertebra centrum identification model is established by collecting lumbar vertebra X-ray images of a lumbar vertebra patient, screening and image processing; acquiring a trabecular bone image from the image of the lumbar vertebra vertebral body recognition model, analyzing morphological parameters and change rules of the trabecular bone, and measuring and evaluating the structural health condition of the trabecular bone; based on the health condition of the trabecular bone structure, images of healthy trabeculae, hidden fractures and trabecular bone injuries in the trabecular bone image are identified, learning and training are carried out on the convolutional neural network through different types of trabecular bone images, and the category of the trabecular bone image is automatically identified through the trained convolutional neural network. By the method, the load of the trabecular bone and the fracture can be judged more accurately, meanwhile, more abundant types of image data can be judged and processed in a large batch, the basic labor of doctors can be greatly released, the diagnosis efficiency is improved, and the medical cost is saved.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
fig. 1 is a schematic flow chart of an artificial risk assessment method for lumbar vertebrae trabecular load stress change and concealed fracture provided by the invention.
Detailed Description
For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
As shown in fig. 1, the present invention provides a lumbar vertebrae trabecular load stress change and concealed fracture artificial risk assessment method, comprising:
collecting lumbar X-ray images of a lumbar patient, screening and processing the images, and establishing a lumbar vertebral body identification model;
acquiring a trabecular bone image from the image of the lumbar vertebra vertebral body recognition model, analyzing morphological parameters and change rules of the trabecular bone, and measuring and evaluating the structural health condition of the trabecular bone;
based on the health condition of the trabecular bone structure, images of healthy trabeculae, hidden fractures and trabecular bone injuries in the trabecular bone image are identified, learning and training are carried out on the convolutional neural network through different types of trabecular bone images, and the category of the trabecular bone image is automatically identified through the trained convolutional neural network.
The invention collects the lumbar vertebra X-ray image sets of 1000 lumbar vertebra patients for analysis, and the inclusion standard comprises the following steps:
the lumbar X-ray image shooting time point is within three months from the current time point, and the corresponding patient does not receive treatment affecting BMD;
the lumbar X-ray image includes a front view and a side view, including at least images of the first to fourth lumbar vertebrae.
Wherein, in the step of screening lumbar vertebrae X-ray image to lumbar vertebrae patient, the exclusion standard includes:
the patient corresponding to the lumbar X-ray image is subjected to lumbar surgery with internal fixation or bone cement filling;
at least one of the first to fourth lumbar vertebrae represents a lesion of a tumor, inflammatory disease, or scoliosis or deformity;
the lumbar X-ray image cannot be matched with the mask;
the resolution of the lumbar X-ray image is low.
For the overall effect of lumbar spine identification, an edge-based segmentation method is employed. An important approach to image segmentation is by edge detection, i.e. detecting where a gray level or structure has a sudden change, indicating the end of one region, and where another region starts. Such discontinuities are referred to as edges. Different images have different gray levels, and the boundary generally has obvious edges, so that the images can be segmented by utilizing the characteristics. The grey value of the pixels at the edges in the image is not continuous and this discontinuity can be detected by taking the derivative. For step-like edges, the position corresponds to the extreme point of the first derivative and to the zero crossing point of the second derivative (zero crossing point). Differential operators are therefore commonly used for edge detection. Because the X-ray image has complex components, unclear boundaries, much noise and high segmentation difficulty, the study on the segmentation problem of the lumbar vertebral body of the X-ray image is rare at present.
With the progress of image technologies such as Micro-CT and the like, morphological parameters of a trabecular bone structure are continuously enriched, a measuring method is expanded from a two-dimensional image to a three-dimensional space, and parameter analysis is converted from qualitative description to quantitative measurement. The morphological parameters of trabeculae mainly comprise thickness, gap, specific surface area, connectivity, structural model index, anisotropy and the like of the trabeculae. Research shows that the morphological parameters of trabeculae present a certain trend along with the change of the position in the bone, the trabeculae at different positions have obviously different structural functions, the condition of the trabeculae at the present stage can be represented and the possible development trend can be predicted by utilizing the morphological parameters and the parameter change condition in a period of time, and therefore, the personalized porous structure is designed according to the relationship between the structure and the function and is applied to replacing or supporting damaged tissues. Hildebrand et al propose a calculation method of structural model index in combination with the composition mode of the trabecular bone structure to describe the change condition of the trabecular bone structure, and draw the conclusion that when a plate-shaped beam is converted into a columnar beam, the structural model index changes remarkably and diseases such as fracture are more likely to occur. Besides the goal of quantitatively expressing the structural state of the trabecula bone, such as thickness, gap and the like, the structural direction, liquid circulation and the like of the trabecula bone are measured, and therefore measurement calculation of connectivity and anisotropy is introduced. Comston proposes an alternative measurement method of "skeletonization" in which only the communication condition is displayed without considering the apparent thickness of the trabecular bone, etc., and finds out the communication change condition of the trabecular bone in the bone. Cook and Larke adopt a more advanced mode factor method to measure the connectivity of the trabecular bone, but as the connectivity of the trabecular bone structure is deeply researched, the mode factor method is more suitable for measuring the tissue morphology of a two-dimensional image, and the relation between the mode factor method and the three-dimensional structure change cannot be established. With the development of three-dimensional stereology technology, a topological tree model with Euler number as a variable is proposed and applied to the quantitative analysis of trabecular bone connectivity. And researching the change condition of trabecular bone connectivity in normal and disease states by using the Euler number to obtain the measurement that the topological tree model is more suitable for three-dimensional connectivity. In addition, the anisotropy quantitatively shows the directionality, the symmetry and the like of the trabecula through the measurement of the external load condition of the trabecula, and has a certain reference value for judging the risk of bone tissue performance reduction caused by the change of the trabecula structure. The anisotropy measurement calculation of the trabecular bone by Williams et al obtains the change trend of the morphological structure of the trabecular bone structure. Lente, Moreno et al measure the degree of anisotropy of trabeculae using the Mean Intercept Length (MIL) algorithm based on two-dimensional images. Through continuous improvement and research of researchers, an MIL algorithm is increased from a two-dimensional image to a three-dimensional structure research.
Clear and divisible trabecular bone images are obtained by utilizing Micro-CT, and two-dimensional image analysis or three-dimensional model analysis can be carried out on the trabecular bone images, so that the change of the trabecular bone can be observed conveniently, and the state of bone tissues can be analyzed. In Micro-CT, based on different experimental purposes, it is necessary to select a proper resolution for image acquisition and analysis, so as to reduce the complexity and error rate of data processing and obtain the best experimental result. The trabecular bone direction in an original three-dimensional reconstruction image directly obtained by Micro-CT scanning reconstruction is not consistent with the coordinate axis direction, the main force line direction is enabled to be parallel to the x, y and z coordinate axis directions through three-dimensional rotation and three-dimensional shearing operation, and data of an interested area is intercepted and converted into a DICOM format file to be output so as to facilitate subsequent structure processing and finite element analysis.
The research on the trabecular bone structure focuses on the research on mechanical properties in the early stage, and the research on the form of the trabecular bone structure is increasingly carried out along with the clear display of the spatial conformation of the trabecular bone. Through continuous trial and discovery of scientific researchers, the morphometric method can measure and calculate the morphological parameters of the trabecula and represent the structural characteristics of the trabecula. Morphometry belongs to the cross-arm of both stereology and biomedical morphometry. The method is characterized in that the tissue structure is observed from a slicing angle and a three-dimensional angle according to the physiological and anatomical basis of trabecula ossis, and quantitative and qualitative description is carried out. The parameter and the parameter change rule are researched by using a form metering method, the possible change rule of the bone trabecular structure state and the load stress can be known, the health condition of the bone trabecular structure can be evaluated, the change of the bone trabecular structure under different loads can be researched and designed, and the change of the structure and the mechanical property of the structure can be found according to the analysis of image textures
Wherein, the morphological parameters of the trabecular bone at least comprise:
bone volume, characterizing trabecular bone volume within the target region;
a total volume, representing the total volume of the target region;
volume fraction, representing the ratio of trabecular bone volume to total volume;
bone surface area, characterizing the surface area of trabeculae of bone in the target region;
specific surface area, which characterizes the total area per unit volume;
trabecular bone thickness, which characterizes the average thickness between beam structures;
trabecular bone gap, representing the average distance between beam structures, when increased, the display distance is increased, the structure is reduced, and the performance is reduced;
the number of trabeculae represents the number of intersection points of the beam units and the non-beam structures, and is related to thickness and gap variation;
the structural model index represents the composition conditions of trabecular bone and columnar beam, when bone pathological changes occur, the trabecular bone is reduced, the columnar trabecular bone is increased, and the structural model index is increased;
connectivity, characterizing the interconnection condition of the beam structure;
anisotropy represents the ratio of the long axis to the short axis in the intercept length fitting ellipsoid in the target region, and the anisotropy value of the load-bearing trabecula bone is increased at the initial stage of bone lesion and is reduced along with the aggravation of the disease.
Research shows that the trabecular bone is an ideal stress model structure, and the beam structure changes along with the change of the external load, namely the trabecular bone structure and the function are kept unified and restricted with each other. Therefore, the trabecular bone is always pressed along the axial direction, the structural morphology of the trabecular bone is consistent with the quantitative distribution of mechanical load through the structural morphology parameter analysis and the bionic reconstruction shrinkage or stretching. However, the structural features of trabecula bone, such as heterogeneity and porosity, make the study of the structural morphology and mechanical properties of trabecula bone difficult. Thus, the study of trabecular bone structures has focused mainly on the following: (1) the function of trabeculae is determined by which factors; (2) how trabeculae contribute to the performance of the external structure; (3) the change in trabecular texture characteristics under different levels of loading.
The trabecular bone is compressed along the axial direction, the structural morphological parameter analysis and the bionic reconstruction shrinkage or stretching of the trabecular bone are carried out, and the structural morphology of the trabecular bone is consistent with the quantitative distribution of mechanical load. After the step of training the convolutional neural network, the method also comprises the step of evaluating the energy efficiency of the network by using an observer operating characteristic curve (ROC curve). The accuracy is reflected by the size of AUC of the area under the ROC curve, and the AUC is less than or equal to 0.8, so that the accuracy is low; 0.8< AUC ≦ 0.9 indicating moderate accuracy; AUC >0.9, indicating high accuracy.
Wherein, the two-dimensional graph of the trabecular bone is processed by a convolutional neural network algorithm to identify displacement, scaling and other form distortion invariance.
Wherein, in the step of performing image processing on the lumbar X-ray image, a trabecular bone image in the lumbar X-ray image is acquired by an image segmentation method.
The intelligent diagnosis process of the occult fracture is a process of simulating a thinking mode and a problem solving strategy of an orthopedics expert during fracture diagnosis by using a computer and automatically giving a diagnosis conclusion and a treatment scheme of the fracture. Since the degree of damage to soft tissues, nerves and blood vessels around the fracture site and other conditions of concurrent damage cannot be obtained in X-ray diagnosis of fractures, comprehensive diagnosis in connection with various data of patients is required to determine a diagnosis conclusion and a treatment plan. Case reasoning is a process of pushing cases to cases, an abstract thinking transformation process is avoided, intermediate links of knowledge transformation are reduced, and adjustment and testing after case matching enable a system to have self-learning capability of solving unknown problems. Therefore, a case reasoning technology which is extremely similar to the thinking method adopted by an orthopedist for diagnosis becomes a key technology for constructing an intelligent fracture diagnosis and treatment system.
In the specific implementation process, for the overall effect of lumbar vertebra identification, an experiment is carried out by adopting an edge-based segmentation method, and the best effect model is obtained through image segmentation model comparison verification.
The change of trabecular bone texture under different loads is realized by comparing the leisure condition of soldiers with the lumbar X-ray image after half hour of training. The method is to compare and analyze X-ray images of 200 military personnel after ordinary physical examination and training with different intensities. The training intensity of a soldier can be divided into: the first-level training is normal intensity, the second-level training is high intensity, the third-level training is overloaded, the fourth-level training is extreme load, and the difference of different degrees of stress change of the bone trabecula of the soldier is displayed through grading of an artificial intelligence algorithm. Based on the imaging technology, the change of the trabecular bone structure is researched, and the lumbar vertebra state analysis has a more positive effect. The two-dimensional graph of the trabecular bone is processed through a convolutional neural network algorithm to identify displacement, scaling and other form distortion invariance, and part of functions are mainly realized by a pooling layer. Since the feature detection layer of CNN learns from the training data, explicit feature extraction is avoided when CNN is used, but learning is implicitly done from the training data; moreover, because the weights of the neurons on the same feature mapping surface are the same, the network can learn in parallel, which is also a great advantage of the convolutional network relative to the network in which the neurons are connected with each other. The convolution neural network has unique superiority in the aspects of voice recognition and image processing by virtue of a special structure with shared local weight, the layout of the convolution neural network is closer to that of an actual biological neural network, the complexity of the network is reduced by virtue of weight sharing, and particularly, the complexity of data reconstruction in the processes of feature extraction and classification is avoided by virtue of the characteristic that an image of a multi-dimensional input vector can be directly input into the network. The method can qualitatively and quantitatively measure and analyze the trabecular bone morphometry parameters, objectively and truly reflect the biological performance of the trabecular bone, judge the degree of lumbar vertebra injury and further prevent the occurrence of lumbar vertebra injury. The current research is mainly to improve the calculation method of the tissue morphometry so as to more accurately analyze the structural state of the trabecular bone. The Convolutional Neural Network (CNN) is a kind of feedforward neural network, and generally includes a data INPUT layer, a convolutional calculation layer, a ReLU activation layer, a pooling layer, and a full connection layer (INPUT-CONV-ReLU-POOL-FC), and is a neural network in which a conventional matrix multiplication operation is replaced by a convolution operation.
The invention has important significance in the aspects of lumbar vertebra trabecular stress change, trabecular bone fracture diagnosis and trabecular bone hidden fracture prevention. Although it is not clear whether changes in trabecular bone properties are initiating factors in the development and progression of osteoarthritis and their role, they are clearly part of important pathological changes. If the stress change of the trabecular bone of the lumbar vertebra and AI follow-up visit and diagnosis and treatment of the fracture of the trabecular bone can be realized through big data acquisition, an early discovery, early treatment and individualized diagnosis and treatment system for the stress change of the trabecular bone of the lumbar vertebra and the fracture of the trabecular bone is formed, the follow-up visit compliance in the treatment process is improved, and the injury of lumbar diseases can be greatly reduced.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. A lumbar vertebrae trabecular load stress change and hidden fracture artificial risk assessment method is characterized by comprising the following steps:
collecting lumbar X-ray images of a lumbar patient, screening and processing the images, and establishing a lumbar vertebral body identification model;
acquiring a trabecular bone image from the image of the lumbar vertebra vertebral body recognition model, analyzing morphological parameters and change rules of the trabecular bone, and measuring and evaluating the structural health condition of the trabecular bone;
based on the health condition of the trabecular bone structure, images of healthy trabeculae, hidden fractures and trabecular bone injuries in the trabecular bone image are identified, learning and training are carried out on the convolutional neural network through different types of trabecular bone images, and the category of the trabecular bone image is automatically identified through the trained convolutional neural network.
2. The method for assessing artificial risk of altered lumbar trabecular load stress and concealed fracture according to claim 1, wherein the step of screening the X-ray images of the lumbar spine of the lumbar patient includes the inclusion criteria of:
the lumbar X-ray image shooting time point is within three months from the current time point, and the corresponding patient does not receive treatment affecting BMD;
the lumbar X-ray image includes a front view and a side view, including at least images of the first to fourth lumbar vertebrae.
3. The method for assessing artificial risk of altered lumbar trabecular load stress and concealed fractures according to claim 1, wherein the step of screening the X-ray images of the lumbar spine of the lumbar patient includes the exclusion criteria of:
the patient corresponding to the lumbar X-ray image is subjected to lumbar surgery with internal fixation or bone cement filling;
at least one of the first to fourth lumbar vertebrae represents a lesion of a tumor, inflammatory disease, or scoliosis or deformity;
the lumbar X-ray image cannot be matched with the mask;
the resolution of the lumbar X-ray image is low.
4. The method for assessing artificial risk of altered lumbar trabecular load stress and concealed fracture according to claim 1, wherein the morphological parameters of trabecular bone include at least:
bone volume, characterizing trabecular bone volume within the target region;
a total volume, representing the total volume of the target region;
volume fraction, representing the ratio of trabecular bone volume to total volume;
bone surface area, characterizing the surface area of trabeculae of bone in the target region;
specific surface area, which characterizes the total area per unit volume;
trabecular bone thickness, which characterizes the average thickness between beam structures;
trabecular bone gap, representing the average distance between beam structures, when increased, the display distance is increased, the structure is reduced, and the performance is reduced;
the number of trabeculae represents the number of intersection points of the beam units and the non-beam structures, and is related to thickness and gap variation;
the structural model index represents the composition conditions of trabecular bone and columnar beam, when bone pathological changes occur, the trabecular bone is reduced, the columnar trabecular bone is increased, and the structural model index is increased;
connectivity, characterizing the interconnection condition of the beam structure;
anisotropy represents the ratio of the long axis to the short axis in the intercept length fitting ellipsoid in the target region, and the anisotropy value of the load-bearing trabecula bone is increased at the initial stage of bone lesion and is reduced along with the aggravation of the disease.
5. The method for assessing artificial risk of load stress changes and concealed fractures of trabecular bone of lumbar vertebrae as claimed in claim 1, wherein the analysis of structural morphological parameters of trabecular bone pressed along the axial direction is consistent with the bionic reconstruction of contraction or extension, and the structural morphology of trabecular bone is consistent with the quantitative distribution of mechanical load.
6. The method for assessing artificial risk of lumbar trabecular load stress alterations and concealed fractures according to claim 1, wherein after the step of training the convolutional neural network, further comprising evaluating network energy efficiency using an observer operating characteristic curve (ROC curve). The accuracy is reflected by the size of AUC of the area under the ROC curve, and the AUC is less than or equal to 0.8, so that the accuracy is low; 0.8< AUC ≦ 0.9 indicating moderate accuracy; AUC >0.9, indicating high accuracy.
7. The method of claim 1, wherein the two-dimensional graph of trabecular bone is processed by a convolutional neural network algorithm to identify displacement, scaling and other forms of distortion invariance.
8. The method for assessing artificial risk of altered lumbar trabecular load stress and concealed fracture according to claim 1, wherein in the step of image processing the lumbar X-ray image, the trabecular image in the lumbar X-ray image is obtained by image segmentation.
9. A computer device comprising an input-output unit, a memory and a processor, wherein the memory has stored therein computer readable instructions which, when executed by the processor, cause the processor to perform the steps of the lumbar trabecular load stress alteration and concealed fracture artificial risk assessment method of any one of claims 1-8.
10. A storage medium storing computer readable instructions which, when executed by one or more processors, cause the one or more processors to perform the steps of the method for artificial risk assessment of lumbar trabecular load stress alteration and occult fracture as claimed in any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010769350.XA CN111899880A (en) | 2020-08-03 | 2020-08-03 | Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010769350.XA CN111899880A (en) | 2020-08-03 | 2020-08-03 | Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111899880A true CN111899880A (en) | 2020-11-06 |
Family
ID=73183245
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010769350.XA Pending CN111899880A (en) | 2020-08-03 | 2020-08-03 | Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111899880A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112396591A (en) * | 2020-11-25 | 2021-02-23 | 暨南大学附属第一医院（广州华侨医院） | Osteoporosis intelligent evaluation method based on lumbar X-ray image |
| CN113240661A (en) * | 2021-05-31 | 2021-08-10 | 平安科技（深圳）有限公司 | Deep learning-based lumbar vertebra analysis method, device, equipment and storage medium |
| WO2023010248A1 (en) * | 2021-08-02 | 2023-02-09 | 香港中文大学 | Apparatus for examining osteoporotic vertebral fracture by using thoracoabdominal frontal view radiograph |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108491770A (en) * | 2018-03-08 | 2018-09-04 | 李书纲 | A kind of data processing method based on fracture image |
| US20200160515A1 (en) * | 2018-11-21 | 2020-05-21 | Siemens Healthcare Gmbh | Processing a Medical Image |
| CN111417980A (en) * | 2017-12-01 | 2020-07-14 | Ucb生物制药有限责任公司 | Three-dimensional medical image analysis method and system for identification of vertebral fractures |
-
2020
- 2020-08-03 CN CN202010769350.XA patent/CN111899880A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111417980A (en) * | 2017-12-01 | 2020-07-14 | Ucb生物制药有限责任公司 | Three-dimensional medical image analysis method and system for identification of vertebral fractures |
| CN108491770A (en) * | 2018-03-08 | 2018-09-04 | 李书纲 | A kind of data processing method based on fracture image |
| US20200160515A1 (en) * | 2018-11-21 | 2020-05-21 | Siemens Healthcare Gmbh | Processing a Medical Image |
| CN111210897A (en) * | 2018-11-21 | 2020-05-29 | 西门子医疗有限公司 | Processing medical images |
Non-Patent Citations (2)
| Title |
|---|
| 彭云等: "CT纹理分析联合机器学习对椎体隐匿性骨折辅助诊断价值初探", 暨南大学学报(自然科学与医学版), vol. 41, no. 03, pages 268 - 274 * |
| 郭新路: "骨小梁结构形态学参数分析与仿生重建", pages 1 - 11 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112396591A (en) * | 2020-11-25 | 2021-02-23 | 暨南大学附属第一医院（广州华侨医院） | Osteoporosis intelligent evaluation method based on lumbar X-ray image |
| CN113240661A (en) * | 2021-05-31 | 2021-08-10 | 平安科技（深圳）有限公司 | Deep learning-based lumbar vertebra analysis method, device, equipment and storage medium |
| CN113240661B (en) * | 2021-05-31 | 2023-09-26 | 平安科技（深圳）有限公司 | Deep learning-based lumbar vertebra bone analysis method, device, equipment and storage medium |
| WO2023010248A1 (en) * | 2021-08-02 | 2023-02-09 | 香港中文大学 | Apparatus for examining osteoporotic vertebral fracture by using thoracoabdominal frontal view radiograph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111899880A (en) | Lumbar vertebra trabecular load stress change and hidden fracture artificial risk assessment method | |
| CN105009174B (en) | Method and system for assessing tissue fibrosis | |
| CN109559292A (en) | Multi-modality images fusion method based on convolution rarefaction representation | |
| CN110025312B (en) | Method and system for predicting curative effect of herpetic neuralgia based on structural magnetic resonance | |
| CN109480780A (en) | A kind of cerebral apoplexy early warning system and method | |
| Kim et al. | Development of an automatic muscle atrophy measuring algorithm to calculate the ratio of supraspinatus in supraspinous fossa using deep learning | |
| Qureshi | Glaucoma detection in retinal images using image processing techniques: a survey | |
| CN111951219B (en) | Thyroid eye disease screening method, system and equipment based on orbit CT image | |
| Sezer et al. | Convolutional neural network based diagnosis of bone pathologies of proximal humerus | |
| CN110148108A (en) | Herpes zoster neuralgia curative effect prediction method and system based on functional MRI | |
| CN108695000A (en) | Thyroid gland Diffuse Diseases intelligent diagnosing method based on ultrasonoscopy and system | |
| Wismüller et al. | Introducing anisotropic minkowski functionals and quantitative anisotropy measures for local structure analysis in biomedical imaging | |
| CN114947756A (en) | Atopic dermatitis severity intelligent evaluation decision-making system based on skin image | |
| CN117153379B (en) | Prediction device for thoracic outlet syndrome | |
| CN111062936A (en) | Quantitative index evaluation method for facial deformation diagnosis and treatment effect | |
| CN114176616A (en) | Venous thrombosis detection method, electronic device and storage medium | |
| Goudarzi et al. | Segmentation of arm ultrasound images in breast cancer-related lymphedema: A database and deep learning algorithm | |
| CN107194417A (en) | Injury of finger detection method and system based on MR images | |
| TWI399194B (en) | Semi-automatic knee cartilage mri image segmentation based on cellular automata | |
| CN110459303A (en) | Medical imaging abnormal detector based on depth migration | |
| Zhang et al. | [Retracted] Deconstruction of Knee Cartilage Injury in Athletes Using MR Images Based on Artificial Intelligence Segmentation Algorithm | |
| Perumal et al. | Tumor Identification in CTScan/UltraSound Using Deep Learning Technique | |
| Yankovy et al. | Classifier of liver diseases according to textural statistics of ultrasound investigation and convolutional neural network | |
| Hernandez et al. | Image analysis tool with laws' masks to bone texture | |
| Amiya et al. | Assertion of Low Bone Mass in Osteoporotic X-ray images using Deep Learning Technique |
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 |