CN112907563A - CTA full-automatic collateral circulation scoring method and system based on deep learning - Google Patents
CTA full-automatic collateral circulation scoring method and system based on deep learning Download PDFInfo
- Publication number
- CN112907563A CN112907563A CN202110286102.4A CN202110286102A CN112907563A CN 112907563 A CN112907563 A CN 112907563A CN 202110286102 A CN202110286102 A CN 202110286102A CN 112907563 A CN112907563 A CN 112907563A
- Authority
- CN
- China
- Prior art keywords
- cta
- brain
- image
- scoring
- map
- 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.)
- Granted
Links
- 206010069729 Collateral circulation Diseases 0.000 title claims abstract description 51
- 238000013135 deep learning Methods 0.000 title claims abstract description 24
- 238000013077 scoring method Methods 0.000 title claims abstract description 19
- 210000004556 brain Anatomy 0.000 claims abstract description 72
- 210000004204 blood vessel Anatomy 0.000 claims abstract description 39
- 239000013598 vector Substances 0.000 claims abstract description 39
- 230000011218 segmentation Effects 0.000 claims abstract description 22
- 238000010606 normalization Methods 0.000 claims abstract description 21
- 230000017531 blood circulation Effects 0.000 claims abstract description 12
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 10
- 238000013145 classification model Methods 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims abstract description 6
- 230000000873 masking effect Effects 0.000 claims abstract description 6
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 19
- 210000001627 cerebral artery Anatomy 0.000 claims description 18
- 210000004720 cerebrum Anatomy 0.000 claims description 17
- 230000002490 cerebral effect Effects 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 210000003625 skull Anatomy 0.000 claims description 6
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 5
- 238000002224 dissection Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 230000001131 transforming effect Effects 0.000 claims description 5
- 230000002792 vascular Effects 0.000 claims description 5
- 230000011157 brain segmentation Effects 0.000 claims description 4
- 230000004087 circulation Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 230000006835 compression Effects 0.000 claims description 3
- 238000007906 compression Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 3
- 238000010968 computed tomography angiography Methods 0.000 abstract description 66
- 238000011156 evaluation Methods 0.000 abstract description 5
- 238000010969 brain CTA Methods 0.000 abstract description 3
- 230000007246 mechanism Effects 0.000 abstract description 3
- 238000012545 processing Methods 0.000 abstract description 2
- 238000013459 approach Methods 0.000 abstract 1
- 230000000302 ischemic effect Effects 0.000 description 9
- 208000032382 Ischaemic stroke Diseases 0.000 description 8
- 238000012163 sequencing technique Methods 0.000 description 7
- 208000006011 Stroke Diseases 0.000 description 6
- 239000008280 blood Substances 0.000 description 6
- 210000004369 blood Anatomy 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 230000002537 thrombolytic effect Effects 0.000 description 4
- 238000011282 treatment Methods 0.000 description 4
- 239000003146 anticoagulant agent Substances 0.000 description 3
- 210000001367 artery Anatomy 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000036770 blood supply Effects 0.000 description 3
- 206010008118 cerebral infarction Diseases 0.000 description 3
- 208000026106 cerebrovascular disease Diseases 0.000 description 3
- 238000007917 intracranial administration Methods 0.000 description 3
- 210000003657 middle cerebral artery Anatomy 0.000 description 3
- 210000001519 tissue Anatomy 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 206010053648 Vascular occlusion Diseases 0.000 description 2
- 230000001154 acute effect Effects 0.000 description 2
- 230000003872 anastomosis Effects 0.000 description 2
- 230000003727 cerebral blood flow Effects 0.000 description 2
- 239000002872 contrast media Substances 0.000 description 2
- 230000002008 hemorrhagic effect Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000010412 perfusion Effects 0.000 description 2
- 210000003446 pia mater Anatomy 0.000 description 2
- 210000003388 posterior cerebral artery Anatomy 0.000 description 2
- 230000010410 reperfusion Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 238000002560 therapeutic procedure Methods 0.000 description 2
- 208000021331 vascular occlusion disease Diseases 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 208000031104 Arterial Occlusive disease Diseases 0.000 description 1
- 201000001320 Atherosclerosis Diseases 0.000 description 1
- 206010008190 Cerebrovascular accident Diseases 0.000 description 1
- 208000005189 Embolism Diseases 0.000 description 1
- 238000012276 Endovascular treatment Methods 0.000 description 1
- 208000016988 Hemorrhagic Stroke Diseases 0.000 description 1
- 206010033799 Paralysis Diseases 0.000 description 1
- 208000007536 Thrombosis Diseases 0.000 description 1
- 206010057469 Vascular stenosis Diseases 0.000 description 1
- 210000002551 anterior cerebral artery Anatomy 0.000 description 1
- 208000021328 arterial occlusion Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000740 bleeding effect Effects 0.000 description 1
- 230000008081 blood perfusion Effects 0.000 description 1
- 210000005013 brain tissue Anatomy 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 210000000269 carotid artery external Anatomy 0.000 description 1
- 210000004004 carotid artery internal Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 210000000275 circle of willis Anatomy 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000003527 fibrinolytic agent Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 208000034783 hypoesthesia Diseases 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 238000012905 input function Methods 0.000 description 1
- 208000020658 intracerebral hemorrhage Diseases 0.000 description 1
- 238000001990 intravenous administration Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000241 respiratory effect Effects 0.000 description 1
- 238000011268 retreatment Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
- 208000011580 syndromic disease Diseases 0.000 description 1
- 229960000103 thrombolytic agent Drugs 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/002—Image coding using neural networks
-
- 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
- G06T2207/20076—Probabilistic image processing
-
- 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
- G06T2207/30016—Brain
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
- G06T2207/30104—Vascular flow; Blood flow; Perfusion
-
- 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
- G06T2207/30204—Marker
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention provides a CTA full-automatic collateral circulation scoring method and a CTA full-automatic collateral circulation scoring system based on deep learning, which relate to the technical field of medical image processing and comprise the following steps: step S1: preprocessing the brain CTA image by masking and normalization; step S2: dividing brain regions of the CTA image, obtaining a related anatomical map and a related functional map, and obtaining mask maps of all the brain regions of the weighted brain; step S3: performing blood vessel segmentation on the mask and the normalized CTA image; step S4: quantitatively calculating scoring characteristics based on the blood vessel segmentation result; step S5: measuring scoring characteristics of the thickness of the blood vessel wall based on the convolutional neural network; step S6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation. The invention can realize a collateral circulation blood vessel grading mechanism based on a blood flow compensation approach, improve the segmentation precision of small blood vessels, and enable a scoring strategy to have wider application range and more objective evaluation threshold.
Description
Technical Field
The invention relates to the technical field of medical image processing, in particular to a CTA full-automatic collateral circulation scoring method and system based on deep learning.
Background
Stroke is currently the leading cause of acquired disability in patients and is also one of the leading causes of death worldwide. Most of cerebral apoplexy is ischemic stroke, mainly cerebral infarction, and vascular stenosis and occlusion caused by intracranial and extracranial atherosclerosis are the most common causes. Stroke is a common disease in clinic, has acute onset and obvious symptoms, and is mainly manifested by paralysis and numbness of limbs. Many patients, despite the prescribed treatment, have difficulties in avoiding more or less sequelae, which is an important threat to human health. The stroke has high treatment cost, and the life quality is influenced, thereby bringing great economic pressure to patients and society.
In ischemic cerebrovascular disease, arterial occlusion results in a decrease in perfusion in the area of the occluded vessel supplying blood. If the blood flow perfusion is <10 ml/(100 g.min), the nerve cells will be irreversibly damaged within a few minutes. If only a fraction of Cerebral Blood Flow (CBF) is reduced, nerve cells will cease to function, but remain structurally intact, a potential tissue that is considered to be salvageable (ischemic penumbra), making them the target of thrombolytic therapy. Collateral circulation of the brain refers to the condition that when the blood supply artery of the brain is severely narrowed or blocked, blood reaches ischemic tissues through other branch blood vessels, so that the ischemic areas obtain different degrees of blood perfusion compensation. Studies have shown that patients with good collateral circulation have larger ischemic penumbra, smaller cerebral infarction area, better thrombolysis and retreatment results. Therefore, accurate assessment of the collateral circulation of the brain is an important condition for the correct treatment of patients with Acute Ischemic Stroke (AIS).
CT Angiography (CTA) is the most common technique for evaluating the intracranial vascular system, has the advantages of rapid imaging, no wound, simple image post-processing and the like, can be used for determining the specific position of vascular occlusion, and is the preferred imaging examination for emergency patients. With this technique, the progress of the contrast agent through the brain can be recorded by acquiring dynamic images. In order to accurately assess the cerebral collateral circulation of a patient, the Arterial Input Function (AIF) and Venous Output Function (VOF) in CTA images need to be measured as the Concentration-Time curves (AUC) of Arterial and Venous voxels containing 100% blood, respectively.
The current research on collateral circulation scoring systems is mainly focused on two areas: a coarse visual score and a quantitative threshold scoring system.
The Early CT Score (ASPECTS) of the Alberta Stroke Program Early CT Score is a grading system for evaluating acute ischemic Stroke CT ischemic change, can evaluate the Early lesion part of ischemic Stroke, and is currently used for a plurality of clinical tests of intravascular mechanical thrombus removal. But because the ASPECT score is judged by people, the ASPECT score is easily influenced by subjective factors of evaluators.
Maas score lesion side branch vessels were compared to normal side branch vessels from CTA raw images (CTA-SI) and scored. The scoring method was used to assess the presence or absence of and grade collateral blood vessels in the lateral fissure region and the pia mater bulge region. But in fact about one quarter of patients cannot measure substantial collateral circulation on CTA through the collateral fissure and pia mater bulge vessels.
The Tan score is based on the collateral vessel filling score of the middle cerebral artery infarct ischemic area, and the score: 3 points (collateral filling up to 100%), 2 points (collateral filling up to 50% ~ 100%), 1 point (0< collateral filling up to < 50%), 0 points (no collateral blood supply). Poor side branch circulation is rated at 0-1, and good side branch circulation is rated at 2-3. The scoring threshold for this score was determined by clinical studies and was only applicable to patients who were intravenously injected with rt-PA thrombolytic agent or who did not receive thrombolytic therapy. For those patients who use other stroke interventional regimens, there may be different thresholds, and thus the scoring system has some limitations in repeatability.
Visual scoring systems can be subject to subjective interpretation, resulting in inter-observer and intra-observer variability. The scoring threshold of the scoring system quantifying the evaluation criteria is susceptible to variation between different samples, and the automatic scoring system can promote objective and repeatable assessment of the lateral dendritic state of the brain through deep learning.
Collateral circulation scores were dependent on the difference between the area of the Middle Cerebral Artery (MCA) on the occluded side and the arterial tree on the contralateral side. Vessel segmentation is therefore an essential step in the realization step. At present, a Convolutional Neural Network (CNN) shows excellent performance in many medical vessel image segmentation tasks, which makes implementation of full-automatic collateral circulation scoring possible. Based on the deep learning method, the difference between the arterial trees on the affected side and the healthy side is quantified, and the efficacy and efficiency of the endovascular treatment and diagnosis opinion on stroke can be effectively improved.
The prior document Su Jianang, et al, "Automatic clinical Scoring From 3D CTA images," IEEE transactions on medical imaging 39.6(2020), wherein Su et al studies this direction and proposes to use 3D U-net to segment blood vessels and use a machine learning classifier to infer the segmentation result to score a threshold in a Tan Scoring system. However, this method assumes that the cause of ischemic stroke is the occurrence of large vessel embolism in the cerebral hemisphere, and therefore the scoring system is not applicable to patients who do not meet this assumption, and the patient population is studied to account for 14-30% of the population with ischemic stroke.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a CTA full-automatic collateral circulation scoring method and system based on deep learning.
According to the CTA full-automatic collateral circulation scoring method and system based on deep learning provided by the invention, the scheme is as follows:
in a first aspect, a CTA full-automatic collateral circulation scoring method based on deep learning is provided, and the method includes:
step S1: acquiring a cerebral CTA image and a corresponding binary mask data set, acquiring a cerebral artery probability density map and a cerebral hemisphere map, and performing mask and normalization preprocessing on the cerebral CTA image;
step S2: dividing brain regions of the CTA image, obtaining a related anatomical map and a related functional map, and obtaining mask maps of all the brain regions of the weighted brain;
step S3: performing blood vessel segmentation on the mask and the normalized CTA image;
step S4: quantitatively calculating scoring characteristics based on the blood vessel segmentation result;
step S5: measuring scoring characteristics of the thickness of the blood vessel wall based on the convolutional neural network;
step S6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation.
Preferably, the step S1 includes:
performing skull dissection and midline correction on the data set;
registering the manually marked data sets using continuous affine and deformable symmetric differential isomorphic image registration;
carrying out spatial normalization, and obtaining a cerebral artery probability density map through B-spline interpolation transformation;
defining and manually marking a cerebral hemisphere map in CTA image space;
correspondingly transforming the binary brain mask, the cerebral artery probability density map and the hemisphere map into each CTA image space for masking;
CTA images were normalized using maximum-minimum normalization.
Preferably, the step S2 includes:
carrying out anatomical atlas brain region division on the CTA image;
performing functional map brain area division on the CTA image;
carrying out weight assignment on each brain area;
CTA images after weighted brain segmentation are obtained.
Preferably, the step S3 includes:
obtaining a pre-processed CTA image;
expanding each manually marked single-pixel center line by using a 3 multiplied by 3 square structural element;
inputting a training image into a network model, the model comprising an encoder and a decoder;
the encoder processes the image, and each step of operation comprises two times of 3 × 3 × 3 convolution for extracting features and one time of 2 × 2 × 2 pooling for down-sampling;
the decoder processes the image, each step comprises deconvolution of 3 multiplied by 3 to carry out up-sampling, and a feature map with the same size from a compression path is spliced, and finally, two times of convolution are carried out;
measuring the performance of the network model by using the dice score, and training;
inputting a test image into the network model;
the network model outputs the segmentation results of the test CTA image.
Preferably, the step S4 includes:
generating an affected side binary chart and an uninjured side binary chart based on clinical information and a cerebral hemisphere chart;
the vessel volume ratio in the volume plane and the maximum phase time are calculated and weighted:
where p represents a voxel in the brain, HYIs the affected side of the brain, HNIs healthy side brain, W (D)i) Weight values representing tertiary vessel compartmentalization, TYTime phase time, T, representing the maximum volume value obtained at the affected sideNRepresenting the time phase when the healthy side obtains the maximum volume value;
calculating the blood flow rate in the flow rate direction;
wherein, VYAnd VNRepresenting the vascular flow rates on the affected and healthy sides, respectively.
Preferably, the step S5 includes:
inputting a CTA image of the newly built collateral circulation vessel, and cutting the picture into 16 patches;
each patch passes through 5 convolution layers and 1 full-connection layer, and after each convolution layer, batch normalization and parametric correction linear unit activation are carried out;
the full connection layers of all the patch are connected through an additional full connection layer;
outputting through a linear activation output layer;
calculating an adaptive learning rate by using an adaptive moment estimation optimizer and training a model;
inputting a lateral branch circulation CTA image to be measured;
outputting the measured value r of the thickness of the collateral circulation vessels;
Preferably, the step S6 includes:
fusing all the characteristic values to construct a characteristic vector f;
calculating the median of each image feature vector f;
inputting a median value of the feature vector f, and determining a classification threshold value according to a Tan scoring standard and clinical experience;
training a support vector classifier, and determining an optimal threshold value of a median value of a feature vector f;
inputting a complete feature vector f, and training a random forest classifier;
training and inputting an ordered regression of the complete feature vectors;
performing ensemble learning on a plurality of models;
testing a grading threshold value obtained by ensemble learning;
screening a scoring threshold range with high accuracy and sensitivity;
and determining a brain collateral scoring standard, and performing classification scoring on the collateral circulation CTA image.
In a second aspect, a CTA full-automatic collateral circulation scoring system based on deep learning is provided, the system comprising:
module M1: acquiring a cerebral CTA image and a corresponding binary mask data set, acquiring a cerebral artery probability density map and a cerebral hemisphere map, and performing mask and normalization preprocessing on the cerebral CTA image;
module M2: dividing brain regions of the CTA image, obtaining a related anatomical map and a related functional map, and obtaining mask maps of all the brain regions of the weighted brain;
module M3: performing blood vessel segmentation on the mask and the normalized CTA image;
module M4: quantitatively calculating scoring characteristics based on the blood vessel segmentation result;
module M5: measuring scoring characteristics of the thickness of the blood vessel wall based on the convolutional neural network;
module M6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation.
Preferably, the module M1 includes:
performing skull dissection and midline correction on the data set;
registering the manually marked data sets using continuous affine and deformable symmetric differential isomorphic image registration;
carrying out spatial normalization, and obtaining a cerebral artery probability density map through B-spline interpolation transformation;
defining and manually marking a cerebral hemisphere map in CTA image space;
correspondingly transforming the binary brain mask, the cerebral artery probability density map and the hemisphere map into each CTA image space for masking;
CTA images were normalized using maximum-minimum normalization.
Preferably, the module M2 includes:
carrying out anatomical atlas brain region division on the CTA image;
performing functional map brain area division on the CTA image;
carrying out weight assignment on each brain area;
CTA images after weighted brain segmentation are obtained.
Compared with the prior art, the invention has the following beneficial effects:
1. according to the invention, the brain functional region is divided, and different regions are weighted, so that a collateral circulation blood vessel grading mechanism based on a blood flow compensation path is realized, and the small blood vessel segmentation precision is improved;
2. according to the method, the blood vessel volume, the weighted volume and the blood vessel flow rate are taken as characteristics to obtain the volume and flow rate ratio of the affected side and the healthy side in each stage, so that the scoring strategy has a wider application range and a more objective evaluation threshold value.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
fig. 1 is a flowchart of a method for providing a brain collateral circulation scoring based on deep learning according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a U-net network according to an embodiment of the present invention;
fig. 3 is a flow chart illustrating a method for scoring collateral circulation of brain according to an embodiment of the present invention.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
The embodiment of the invention provides a CTA full-automatic collateral circulation scoring method based on deep learning, which is shown in figure 1 and specifically comprises the following steps:
step S1: the method comprises the steps of obtaining a cerebral CTA image and a corresponding binary mask data set, obtaining a cerebral artery probability density map and a cerebral hemisphere map, and conducting mask and normalization preprocessing on the cerebral CTA image. Specifically, the dataset is first subjected to skull dissection and midline correction, and the manually marked dataset is registered using continuous affine and deformable symmetric differential homomorphic image registration. Secondly, space normalization is carried out, a cerebral artery probability density map is obtained through B-spline interpolation transformation, a cerebral hemisphere map is defined and manually marked in a CTA image space, a binary brain mask, the cerebral artery probability density map and the hemisphere map are correspondingly transformed to each CTA image space to be masked, and finally, the CTA image is normalized by using maximum and minimum normalization. The B-spline interpolation described in this embodiment is very widely applied, where the functional properties are also symmetric, and the coefficient matrix is easily obtained by matrix inversion.
Step S2: the CTA image is subjected to brain area division, an Automatic Anatomical Labeling (AAL) anatomical map, an Eichhoff-Zilles (EZ) anatomical map, a Talairach and Tournoux (TT) anatomical map, a Harvard-Oxford (HO) anatomical map, a CC200 functional map and a CC400 functional map are obtained, and weighted brain area mask maps are obtained. Specifically, the CTA image is subjected to anatomical atlas brain area division, then the CTA image is subjected to functional atlas brain area division, then weight assignment is carried out on each brain area, and finally the CTA image after weighted brain area division is obtained.
Referring to fig. 2 and 3, step S3: the mask and the normalized CTA image were subjected to vessel segmentation.
Wherein, fig. 2 is a schematic structural diagram of a U-net network model in the present invention, the U-net network model includes: the device comprises an encoder and a decoder, wherein the encoder is mainly a convolutional layer, the decoder is an upsampling layer, a characteristic diagram obtained by each convolutional layer is connected to the corresponding upsampling layer, and a standard Dice score is used as a loss function. The rms propagation optimizer was chosen with an initial learning rate of 0.1, halved every 10 epochs, set to 600. The annotation data set was randomly divided into training and validation data sets in a 3:1 ratio.
Specifically, a preprocessed CTA image is obtained, each manually labeled single-pixel center line is expanded by using a 3 × 3 × 3 square structural element, and a training image is input into a network model. The encoder processes the image, with each step of operation comprising two 3 × 3 × 3 convolutions to extract features and one 2 × 2 × 2 pooling for down sampling (Downsampling); the decoder processes the image, each step comprising a 3 × 3 × 3 deconvolution for Upsampling (Upsampling) and concatenating the same size Feature Map (Feature Map) from the compression path, followed by two final convolutions. And measuring the performance of the network model by using the dice score, training, inputting the test image into the network model, and finally outputting the segmentation result of the test CTA image by the network model.
Step S4: and quantitatively calculating scoring characteristics based on the blood vessel segmentation result. The method comprises the following specific steps:
and generating an affected side binary chart and an uninjured side binary chart based on the clinical information and the cerebral hemisphere chart.
The vessel volume ratio in the volume plane and the maximum phase time are calculated and weighted:
where p represents a voxel in the brain, HYIs the affected side of the brain, HNIs healthy side brain, W (D)i) Weight values representing tertiary vessel compartmentalization, TYTime phase time, T, representing the maximum volume value obtained at the affected sideNRepresenting the time of the phase at which the healthy side has obtained the maximum volume value.
Calculating the blood flow rate in the flow rate direction:
wherein, VYAnd VNRepresenting the vascular flow rates on the affected and healthy sides, respectively.
Step S5: and measuring the scoring characteristic of the thickness of the blood vessel wall based on the convolutional neural network. Specifically, a CTA image of a newly-built collateral circulation vessel is input, a picture is cut into 16 patches, each patch passes through 5 convolutional layers and 1 full-connected layer, Batch Normalization (BN) and parameterized modified linear unit (PReLU) activation are carried out after each convolutional layer, the full-connected layers of the patches are connected, output is carried out through a linear activation output layer through an additional full-connected layer, an adaptive moment estimation (Adam) optimizer is used for calculating an adaptive learning rate, and a model is trained. Inputting a collateral circulation CTA image to be measured, and finally outputting a measured value r of the thickness of a collateral circulation blood vessels。
Step S6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation. The method comprises the following specific steps: fusing all the characteristic values, constructing a characteristic vector f, calculating the median of all the image characteristic vectors f, then inputting the median of the characteristic vector f, and determining a classification threshold value according to Tan scoring standard and clinical experience. Training a support vector classifier, determining an optimal threshold value of a median value of a feature vector f, inputting a complete feature vector f, training a random forest classifier, training and inputting an ordered regression of the complete feature vector, performing ensemble learning on the four models, testing a scoring threshold value obtained by the ensemble learning, screening a scoring threshold value range with high accuracy and sensitivity, finally determining a brain collateral scoring standard, and performing classification scoring on a collateral circulation CTA image.
Next, the present invention will be described in more detail.
The invention provides a CTA full-automatic collateral circulation scoring method based on deep learning, which comprises the following specific schemes:
referring to fig. 1 and 2, the brain CTA image is preprocessed:
acquiring a brain CTA image and a corresponding binary mask dataset; identifying the advanced venous phase 2 after the peak of arteriovenous and venous phases and the peak of venous phases according to AIF and VOF, and performing skull stripping and midline correction; acquiring a cerebral artery probability density map M and a cerebral hemisphere map H; and correspondingly transforming the binary brain mask, the cerebral artery probability density map M and the hemisphere map H into each CTA image space for masking, and normalizing the CTA image by using maximum and minimum normalization.
Brain area partitioning for CTA images:
the brain area is divided according to an anatomical map and a functional map by being manually marked on a CTA axial position image (combined with a coronal vector position) of a healthy population, so that masks of all the brain areas are obtained. Each brain region is given a weight according to the importance degree of each brain region, and the CTA image is divided in a weighted manner by using the weighted brain region mask.
Referring to fig. 2 and 3, the mask and normalized CTA image were subjected to vessel segmentation:
the vessel is segmented hierarchically using a classical 3D U-net network structure with depth of 5, and the model includes an encoder and a decoder. Adding a residual error module in a path of an encoder to avoid gradient explosion caused by the depth of the encoder and obtain deeper characteristic values; in the cascade decoder path, a depth supervision module is added to avoid information loss and gradient disappearance in each convolutional layer. The input data to the model is a 64 × 64 × 64 voxel cube, and a classifier is used to output a vessel classification map U of three-dimensional voxels, and the vessels are classified into class I, class II, and class III branches according to the blood flow compensation pathway. Wherein, the grade I collateral branch is the anastomotic arterial loop (Willis loop, Circle of Willis) located at the bottom of the brain; the II-level side branch comprises an intracranial leptomeningeal artery anastomosis channel between the far ends of the anterior, middle and posterior cerebral arteries, between the posterior cerebral artery and the main cerebellar artery and an anastomosis channel between an external carotid artery branch and an internal carotid artery branch; grade III collateral refers to newly formed capillaries around ischemic tissue.
And (3) quantitatively calculating scoring characteristics based on the blood vessel segmentation result:
the artificial subjective scoring of the cerebral collateral circulation is based on the comparison of the number of visible blood vessels on the affected side and the non-affected side, and the full-automatic scoring method adopts a similar strategy: and (4) combining the blood vessel grading diagram U output by using the U-net model and the corresponding cerebral artery probability density diagram M and cerebral hemisphere diagram H to compare and determine the affected side and the non-affected side of the brain. Based on the clinical symptom known affected hemisphere, based on the information and cerebral hemisphere chart H, generating affected side binary chart HYAnd the non-affected side binary drawing HN. Based on the images, the blood vessel volume, the weighted volume, the phase-time ratio of the maximum volume value period, the blood vessel flow velocity and other specific values are obtained, and the four ratios of the volume, the flow velocity and the like of the affected side and the non-affected side in each period are calculated to serve as scoring characteristics, so that the problem that the scoring characteristics need to be manually obtained in an ASPECTS (advanced acute respiratory syndrome) process is solved:
1. volume: the number of blood vessels on the affected side and the non-affected side is obviously different, and the two-step drawing H is performed on the affected sideYAnd the non-affected side binary drawing HNThe vessel volume ratios for each brain partition (R1-R8) were obtained and ranked. It is assumed that in the unaffected hemisphere, the ratio of the vessel volumes (affected volume divided by unaffected volume) should be a value between 0 (no vessels are visible on the affected side) and 1 (the same number of vessels are visible on both sides). The maximum value of the blood vessel volume of the affected side and the blood vessel volume of the non-affected side are summed, and then r is respectively obtained according to the weighting value of the three-level blood vessel subareasVAnd rw。
Where p represents a voxel in the brain, HYIs the affected side of the brain, HNIs healthy side brain, W (D)i) Weight values representing tertiary vessel compartmentalization, TYTime phase time, T, representing the maximum volume value obtained at the affected sideNRepresenting the time of the phase at which the healthy side has obtained the maximum volume value.
2. Phase stage: CTA requires precise arterial phase and is typically measured using a triggering technique or pre-emphasis. The contrast agent is easy to gather in the blood supply rich area, the maximum value period of the development volume of the affected side and the non-affected side is different due to the blood flow velocity and the like, and the ratio r of the time is usedTAs a feature.
3. Flow rate: there is a significant difference in blood flow rate between the affected and non-affected side. By using the ratio r of blood flow velocity of blood in the affected side to blood flow velocity in the non-affected sidevAs a feature.
Wherein, VYAnd VNRepresenting the vascular flow rates on the affected and healthy sides, respectively.
Based on the evaluation characteristics of the convolutional neural network measured vessel wall thickness:
reperfusion is important for the survival of ischemic penumbra brain tissue, but reperfusion may induce hemorrhagic transformation. Post-thrombolytic Hemorrhagic Transformation (HT) refers to Hemorrhagic stroke that occurs within a period of time after treatment with intravenous thrombolysis in patients with acute ischemic stroke. The newly established collateral circulation vessel wall is not well developed, and bleeding is caused by blood impact during recanalization. The invention measures the thickness of the vessel wall of the collateral circulation through the convolutional neural network and takes the thickness as a characteristic of the collateral circulation score.
The network cuts pictures into 16 patches, each patch passes through 5 convolution layers and 1 full-connection layer, the full-connection layers of the patches are connected and pass through an additional full-connection layer, and finally the output is output through a linear activation output layer. All layers except the output layer were normalized for each Batch of data by Parametric modified Linear Unit (PReLU) activation and Batch Normalization (BN), and an adaptive moment estimation (Adam) optimizer was used to calculate the adaptive learning rate for each parameter.
Based on the network, the thickness of the vessel wall thinner than the voxel can be accurately measured, so that the thickness of the collateral circulation vessel wall is measured, and the thickness characteristic is recorded as rs。
Multi-label scoring classification model:
using the feature vector f ═ rV,rw,rT,rv,rs]As an input prediction Tan score (0,1,2 and 3), four methods for determining a score type threshold value are provided for ensemble learning, namely a clinical definition threshold value, a support vector classifier, a random forest classifier and an ordered regression classifier, and after training is finished, threshold value screening is carried out according to the test result of each classifier.
1. Clinically defined thresholds: the median of the feature vector f is used as input, and the classification threshold is determined by using the definition of the clinical cerebral collateral circulation score. The collateral status is graded according to the standard of Tan scoring based on general clinical knowledge and experience in order to provide a reliable reference standard for collateral scoring for the present invention.
2. Support vector classifier: the median of the feature vector f is input and the best threshold for the median of the feature vector f is determined using a Support Vector Classifier (SVC) with linear kernels. The median classification of the feature vectors is done using interval maximization for SVC. The generalization capability of the learning machine is improved by seeking the minimum structured risk, and the minimization of the experience risk and the confidence range is realized, so that the aim of obtaining a good statistical rule under the condition of less statistical sample quantity is fulfilled.
3. A random forest classifier: the random forest is an algorithm integrating a plurality of trees through the idea of ensemble learning, and has the advantages that input samples with high-dimensional features can be processed, dimension reduction is not needed, the importance of each feature on a classification problem can be evaluated, and therefore a complete feature vector f is used as input.
4. An ordered regression classifier: the feature vector generated by the invention is essentially a sequencing variable, namely, between a continuous variable and a fixed-class variable, and is divided into different classes of relative orders on a measurement level, but is not continuous. If a multi-classification model is used for the sequencing variables, the inherent sequencing of the data is ignored, so that sequencing information is lost, and the statistical efficiency of the statistical result is lost because the sequencing information is missed. If a common linear regression model is used, then treating the sequencing variables as continuous variables can result in artificial information expansion. Therefore, for a sequencing variable, a corresponding model is required to fit its properties in both aspects, the most common method being an ordered regression model. Therefore, ordered-class prediction is performed using ordered regression of the input complete feature vectors.
The embodiment of the invention provides a CTA full-automatic collateral circulation scoring method based on deep learning, which is characterized in that brain functional regions are divided, different partitions are weighted, a collateral circulation blood vessel grading mechanism based on a blood flow compensation path is realized, and the small blood vessel segmentation precision is improved; meanwhile, the volume and the flow rate ratio of the affected side and the healthy side in each stage are obtained by taking the blood vessel volume, the weighted volume and the blood vessel flow rate as characteristics, so that the scoring strategy has a wider application range and a more objective evaluation threshold value.
Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, units provided by the present invention as pure computer readable program code, the system and its various devices, modules, units provided by the present invention can be fully implemented by logically programming method steps in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system and various devices, modules and units thereof provided by the invention can be regarded as a hardware component, and the devices, modules and units included in the system for realizing various functions can also be regarded as structures in the hardware component; means, modules, units for performing the various functions may also be regarded as structures within both software modules and hardware components for performing the method.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.
Claims (10)
1. A CTA full-automatic collateral circulation scoring method based on deep learning is characterized by comprising the following steps:
step S1: acquiring a cerebral CTA image and a corresponding binary mask data set, acquiring a cerebral artery probability density map and a cerebral hemisphere map, and performing mask and normalization preprocessing on the cerebral CTA image;
step S2: dividing brain regions of the CTA image, obtaining a related anatomical map and a related functional map, and obtaining mask maps of all the brain regions of the weighted brain;
step S3: performing blood vessel segmentation on the mask and the normalized CTA image;
step S4: quantitatively calculating scoring characteristics based on the blood vessel segmentation result;
step S5: measuring scoring characteristics of the thickness of the blood vessel wall based on the convolutional neural network;
step S6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation.
2. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S1 comprises:
step S1.1: performing skull dissection and midline correction on the data set;
step S1.2: registering the manually marked data sets using continuous affine and deformable symmetric differential isomorphic image registration;
step S1.3: carrying out spatial normalization, and obtaining a cerebral artery probability density map through B-spline interpolation transformation;
step S1.4: defining and manually marking a cerebral hemisphere map in CTA image space;
step S1.5: correspondingly transforming the binary brain mask, the cerebral artery probability density map and the hemisphere map into each CTA image space for masking;
step S1.6: CTA images were normalized using maximum-minimum normalization.
3. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S2 comprises:
step S2.1: carrying out anatomical atlas brain region division on the CTA image;
step S2.2: performing functional map brain area division on the CTA image;
step S2.3: carrying out weight assignment on each brain area;
step S2.4: CTA images after weighted brain segmentation are obtained.
4. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S3 comprises:
step S3.1: obtaining a pre-processed CTA image;
step S3.2: expanding each manually marked single-pixel center line by using a 3 multiplied by 3 square structural element;
step S3.3: inputting a training image into a network model, the model comprising an encoder and a decoder;
step S3.4: the encoder processes the image, and each step of operation comprises two times of 3 × 3 × 3 convolution for extracting features and one time of 2 × 2 × 2 pooling for down-sampling;
step S3.5: the decoder processes the image, each step comprises deconvolution of 3 multiplied by 3 to carry out up-sampling, and a feature map with the same size from a compression path is spliced, and finally, two times of convolution are carried out;
step S3.6: measuring the performance of the network model by using the dice score, and training;
step S3.7: inputting a test image into the network model;
step S3.8: the network model outputs the segmentation results of the test CTA image.
5. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S4 comprises:
step S4.1: generating an affected side binary chart and an uninjured side binary chart based on clinical information and a cerebral hemisphere chart;
step S4.2: the vessel volume ratio in the volume plane and the maximum phase time are calculated and weighted:
where p represents a voxel in the brain, HYIs the affected side of the brain, HNIs healthy side brain, W (D)i) Weight values representing tertiary vessel compartmentalization, TYTime phase time, T, representing the maximum volume value obtained at the affected sideNRepresenting the time phase when the healthy side obtains the maximum volume value;
step S4.3: calculating the blood flow rate in the flow rate direction;
wherein, VYAnd VNRepresenting the vascular flow rates on the affected and healthy sides, respectively.
6. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S5 comprises:
step S5.1: inputting a CTA image of the newly built collateral circulation vessel, and cutting the picture into 16 patches;
step S5.2: each patch passes through 5 convolution layers and 1 full-connection layer, and after each convolution layer, batch normalization and parametric correction linear unit activation are carried out;
step S5.3: the full connection layers of all the patch are connected through an additional full connection layer;
step S5.3: outputting through a linear activation output layer;
step S5.4: calculating an adaptive learning rate by using an adaptive moment estimation optimizer and training a model;
step S5.5: inputting a lateral branch circulation CTA image to be measured;
step S5.5: outputting the measured value r of the thickness of the collateral circulation vessels。
7. The deep learning-based CTA full-automatic collateral circulation scoring method according to claim 1, wherein the step S6 comprises:
step S6.1: fusing all the characteristic values to construct a characteristic vector f;
step S6.2: calculating the median of each image feature vector f;
step S6.3: inputting a median value of the feature vector f, and determining a classification threshold value according to a Tan scoring standard and clinical experience;
step S6.4: training a support vector classifier, and determining an optimal threshold value of a median value of a feature vector f;
step S6.5: inputting a complete feature vector f, and training a random forest classifier;
step S6.6: training and inputting an ordered regression of the complete feature vectors;
step S6.7: performing ensemble learning on a plurality of models;
step S6.8: testing a grading threshold value obtained by ensemble learning;
step S6.9: screening a scoring threshold range with high accuracy and sensitivity;
step S6.10: and determining a brain collateral scoring standard, and performing classification scoring on the collateral circulation CTA image.
8. A CTA full-automatic collateral circulation scoring system based on deep learning is characterized by comprising:
module M1: acquiring a cerebral CTA image and a corresponding binary mask data set, acquiring a cerebral artery probability density map and a cerebral hemisphere map, and performing mask and normalization preprocessing on the cerebral CTA image;
module M2: dividing brain regions of the CTA image, obtaining a related anatomical map and a related functional map, and obtaining mask maps of all the brain regions of the weighted brain;
module M3: performing blood vessel segmentation on the mask and the normalized CTA image;
module M4: quantitatively calculating scoring characteristics based on the blood vessel segmentation result;
module M5: measuring scoring characteristics of the thickness of the blood vessel wall based on the convolutional neural network;
module M6: and constructing a multi-label grading classification model, and carrying out classification grading on the feature vectors obtained by calculation.
9. The deep learning-based CTA full-automatic collateral circulation scoring system of claim 8, wherein the module M1 comprises:
performing skull dissection and midline correction on the data set;
registering the manually marked data sets using continuous affine and deformable symmetric differential isomorphic image registration;
carrying out spatial normalization, and obtaining a cerebral artery probability density map through B-spline interpolation transformation;
defining and manually marking a cerebral hemisphere map in CTA image space;
correspondingly transforming the binary brain mask, the cerebral artery probability density map and the hemisphere map into each CTA image space for masking;
CTA images were normalized using maximum-minimum normalization.
10. The deep learning-based CTA full-automatic collateral circulation scoring system of claim 8, wherein the module M2 comprises:
carrying out anatomical atlas brain region division on the CTA image;
performing functional map brain area division on the CTA image;
carrying out weight assignment on each brain area;
CTA images after weighted brain segmentation are obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110286102.4A CN112907563B (en) | 2021-03-17 | 2021-03-17 | CTA full-automatic collateral circulation scoring method and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110286102.4A CN112907563B (en) | 2021-03-17 | 2021-03-17 | CTA full-automatic collateral circulation scoring method and system based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112907563A true CN112907563A (en) | 2021-06-04 |
CN112907563B CN112907563B (en) | 2023-12-01 |
Family
ID=76105471
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110286102.4A Active CN112907563B (en) | 2021-03-17 | 2021-03-17 | CTA full-automatic collateral circulation scoring method and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112907563B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113610800A (en) * | 2021-08-05 | 2021-11-05 | 西门子数字医疗科技(上海)有限公司 | Device for assessing collateral circulation, non-diagnostic method and electronic apparatus |
CN113963001A (en) * | 2021-10-25 | 2022-01-21 | 中国人民解放军总医院第一医学中心 | Collateral circulation blood vessel quantification method based on multi-temporal CTA |
CN114119519A (en) * | 2021-11-16 | 2022-03-01 | 高峰 | Collateral circulation assessment method |
CN114332043A (en) * | 2021-12-31 | 2022-04-12 | 深圳市铱硙医疗科技有限公司 | Collateral circulation measuring method, device, equipment and medium based on CT perfusion image |
CN114463288A (en) * | 2022-01-18 | 2022-05-10 | 深圳市铱硙医疗科技有限公司 | Brain medical image scoring method, device, computer equipment and storage medium |
CN117291281A (en) * | 2023-08-29 | 2023-12-26 | 强联智创(北京)科技有限公司 | Method for training evaluation model for evaluating brain collateral circulation and related product |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3316217A1 (en) * | 2016-10-31 | 2018-05-02 | Siemens Healthcare GmbH | Deep learning based bone removal in computed tomography angiography |
US20180240235A1 (en) * | 2017-02-23 | 2018-08-23 | Zebra Medical Vision Ltd. | Convolutional neural network for segmentation of medical anatomical images |
US20180276817A1 (en) * | 2017-03-24 | 2018-09-27 | Umc Utrecht Holding B.V. | Method and System for Assessing Vessel Obstruction Based on Machine Learning |
CN108629784A (en) * | 2018-05-08 | 2018-10-09 | 上海嘉奥信息科技发展有限公司 | A kind of CT image intracranial vessel dividing methods and system based on deep learning |
CN110934608A (en) * | 2019-10-31 | 2020-03-31 | 上海杏脉信息科技有限公司 | Cerebral apoplexy early-stage CTA image evaluation system and method and readable storage medium |
CN111369525A (en) * | 2020-03-02 | 2020-07-03 | 联影智能医疗科技(北京)有限公司 | Image analysis method, apparatus and storage medium |
CN111415361A (en) * | 2020-03-31 | 2020-07-14 | 浙江大学 | Method and device for estimating brain age of fetus and detecting abnormality based on deep learning |
WO2020144374A1 (en) * | 2019-01-11 | 2020-07-16 | Pie Medical Imaging Bv | Methods and systems for dynamic coronary roadmapping |
WO2020172188A1 (en) * | 2019-02-19 | 2020-08-27 | Cedars-Sinai Medical Center | Systems and methods for calcium-free computed tomography angiography |
US20200294241A1 (en) * | 2019-03-12 | 2020-09-17 | The General Hospital Corporation | Automatic segmentation of acute ischemic stroke lesions in computed tomography data |
CN111986101A (en) * | 2020-07-09 | 2020-11-24 | 浙江工业大学 | Cerebrovascular map construction method |
CN112101523A (en) * | 2020-08-24 | 2020-12-18 | 复旦大学附属华山医院 | Stroke risk screening method and system for CBCT (cone beam computed tomography) image cross-modal prediction CTA (computed tomography angiography) image based on deep learning |
CN112419282A (en) * | 2020-11-26 | 2021-02-26 | 中国科学院苏州生物医学工程技术研究所 | Automatic detection method and system for aneurysm in brain medical image |
US10950337B1 (en) * | 2018-11-29 | 2021-03-16 | Heuron Co., Ltd. | Aspect score estimating system and method |
-
2021
- 2021-03-17 CN CN202110286102.4A patent/CN112907563B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3316217A1 (en) * | 2016-10-31 | 2018-05-02 | Siemens Healthcare GmbH | Deep learning based bone removal in computed tomography angiography |
US20180240235A1 (en) * | 2017-02-23 | 2018-08-23 | Zebra Medical Vision Ltd. | Convolutional neural network for segmentation of medical anatomical images |
US20180276817A1 (en) * | 2017-03-24 | 2018-09-27 | Umc Utrecht Holding B.V. | Method and System for Assessing Vessel Obstruction Based on Machine Learning |
CN108629784A (en) * | 2018-05-08 | 2018-10-09 | 上海嘉奥信息科技发展有限公司 | A kind of CT image intracranial vessel dividing methods and system based on deep learning |
US10950337B1 (en) * | 2018-11-29 | 2021-03-16 | Heuron Co., Ltd. | Aspect score estimating system and method |
WO2020144374A1 (en) * | 2019-01-11 | 2020-07-16 | Pie Medical Imaging Bv | Methods and systems for dynamic coronary roadmapping |
WO2020172188A1 (en) * | 2019-02-19 | 2020-08-27 | Cedars-Sinai Medical Center | Systems and methods for calcium-free computed tomography angiography |
US20200294241A1 (en) * | 2019-03-12 | 2020-09-17 | The General Hospital Corporation | Automatic segmentation of acute ischemic stroke lesions in computed tomography data |
CN110934608A (en) * | 2019-10-31 | 2020-03-31 | 上海杏脉信息科技有限公司 | Cerebral apoplexy early-stage CTA image evaluation system and method and readable storage medium |
CN111369525A (en) * | 2020-03-02 | 2020-07-03 | 联影智能医疗科技(北京)有限公司 | Image analysis method, apparatus and storage medium |
CN111415361A (en) * | 2020-03-31 | 2020-07-14 | 浙江大学 | Method and device for estimating brain age of fetus and detecting abnormality based on deep learning |
CN111986101A (en) * | 2020-07-09 | 2020-11-24 | 浙江工业大学 | Cerebrovascular map construction method |
CN112101523A (en) * | 2020-08-24 | 2020-12-18 | 复旦大学附属华山医院 | Stroke risk screening method and system for CBCT (cone beam computed tomography) image cross-modal prediction CTA (computed tomography angiography) image based on deep learning |
CN112419282A (en) * | 2020-11-26 | 2021-02-26 | 中国科学院苏州生物医学工程技术研究所 | Automatic detection method and system for aneurysm in brain medical image |
Non-Patent Citations (3)
Title |
---|
RAHIL SHAHZAD ET.AL: "Fully automated detection and segmentation of intracranial aneurysms in subarachnoid hemorrhage on CTA using deep learning", 《NATURE》 * |
刘国玮: "基于深度学习的脑部CTA图像血管分割方法研究", 《中国优秀硕士学位论文全文数据库(信息科技辑)》, no. 03 * |
吴秋雯等: "基于深度学习的计算机体层摄影血管造影颈动脉斑块分割初步研究", 《上海医学》, vol. 43, no. 05, pages 280 - 283 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113610800A (en) * | 2021-08-05 | 2021-11-05 | 西门子数字医疗科技(上海)有限公司 | Device for assessing collateral circulation, non-diagnostic method and electronic apparatus |
CN113963001A (en) * | 2021-10-25 | 2022-01-21 | 中国人民解放军总医院第一医学中心 | Collateral circulation blood vessel quantification method based on multi-temporal CTA |
CN113963001B (en) * | 2021-10-25 | 2023-05-23 | 中国人民解放军总医院第一医学中心 | Side branch circulation vessel quantification method based on multi-phase CTA |
CN114119519A (en) * | 2021-11-16 | 2022-03-01 | 高峰 | Collateral circulation assessment method |
CN114332043A (en) * | 2021-12-31 | 2022-04-12 | 深圳市铱硙医疗科技有限公司 | Collateral circulation measuring method, device, equipment and medium based on CT perfusion image |
CN114463288A (en) * | 2022-01-18 | 2022-05-10 | 深圳市铱硙医疗科技有限公司 | Brain medical image scoring method, device, computer equipment and storage medium |
CN114463288B (en) * | 2022-01-18 | 2023-01-10 | 深圳市铱硙医疗科技有限公司 | Brain medical image scoring method and device, computer equipment and storage medium |
CN117291281A (en) * | 2023-08-29 | 2023-12-26 | 强联智创(北京)科技有限公司 | Method for training evaluation model for evaluating brain collateral circulation and related product |
Also Published As
Publication number | Publication date |
---|---|
CN112907563B (en) | 2023-12-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112907563A (en) | CTA full-automatic collateral circulation scoring method and system based on deep learning | |
Chen et al. | 3D intracranial artery segmentation using a convolutional autoencoder | |
Su et al. | Automatic collateral scoring from 3D CTA images | |
US20220284584A1 (en) | Computerised tomography image processing | |
CN110415234A (en) | Brain tumor dividing method based on multi-parameter magnetic resonance imaging | |
Fan et al. | An automatic estimation of arterial input function based on multi-stream 3D CNN | |
Yalçın et al. | Brain stroke classification and segmentation using encoder-decoder based deep convolutional neural networks | |
CN113223015A (en) | Vascular wall image segmentation method, device, computer equipment and storage medium | |
Hepp et al. | Fully automated segmentation and shape analysis of the thoracic aorta in non–contrast-enhanced magnetic resonance images of the German National Cohort Study | |
CN115605961A (en) | System and method for predicting ischemic brain tissue regression from multi-stage CT angiography of acute ischemic stroke patient by utilizing machine learning | |
Chen et al. | Generative adversarial network based cerebrovascular segmentation for time-of-flight magnetic resonance angiography image | |
Tomasetti et al. | Machine learning algorithms versus thresholding to segment ischemic regions in patients with acute ischemic stroke | |
Gayathri et al. | Fetal brain segmentation using improved maximum entropy threshold | |
CN113939224B (en) | Vertebral artery interlayer risk assessment method, computer equipment and storage medium | |
AU3720700A (en) | 3-d shape measurements using statistical curvature analysis | |
Banerjee et al. | Segmentation of intracranial aneurysm remnant in MRA using dual-attention atrous net | |
Saygılı et al. | Exploiting lamina terminalis appearance and motion in prediction of hydrocephalus using convolutional LSTM network | |
Kim et al. | A single-stage detector of cerebral microbleeds using 3D feature fused region proposal network (FFRP-Net) | |
Autrusseau et al. | Toward a 3d arterial tree bifurcation model for intra-cranial aneurysm detection and segmentation | |
Wang et al. | A 3D cross-hemisphere neighborhood difference Convnet for chronic stroke lesion segmentation | |
Payne et al. | Review of in silico models of cerebral blood flow in health and pathology | |
Mahyoub et al. | Brain Tumor Segmentation in Fluid-Attenuated Inversion Recovery Brain MRI using Residual Network Deep Learning Architectures | |
KR102655827B1 (en) | Brain tissue segmentation System and method from the brain MRI data using deep-learning | |
Lin et al. | StenUNet: Automatic Stenosis Detection from X-ray Coronary Angiography | |
Pal et al. | Segmentation of major cerebral vessel from MRA images and evaluation using U-Net family |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |