CN116705296A - Method and system for risk stratification of GBM patient based on conventional MRI sequence - Google Patents
Method and system for risk stratification of GBM patient based on conventional MRI sequence Download PDFInfo
- Publication number
- CN116705296A CN116705296A CN202310667266.0A CN202310667266A CN116705296A CN 116705296 A CN116705296 A CN 116705296A CN 202310667266 A CN202310667266 A CN 202310667266A CN 116705296 A CN116705296 A CN 116705296A
- Authority
- CN
- China
- Prior art keywords
- image
- histology
- conventional mri
- sequence
- features
- 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
- 208000005017 glioblastoma Diseases 0.000 title claims abstract description 129
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000013517 stratification Methods 0.000 title claims abstract description 50
- 108090000623 proteins and genes Proteins 0.000 claims abstract description 39
- 238000012163 sequencing technique Methods 0.000 claims abstract description 36
- 238000004393 prognosis Methods 0.000 claims abstract description 30
- 238000010199 gene set enrichment analysis Methods 0.000 claims abstract description 13
- 238000013427 histology analysis Methods 0.000 claims abstract description 13
- 238000003012 network analysis Methods 0.000 claims abstract description 13
- 230000004186 co-expression Effects 0.000 claims abstract description 12
- 238000003384 imaging method Methods 0.000 claims description 90
- 238000012216 screening Methods 0.000 claims description 41
- 206010028980 Neoplasm Diseases 0.000 claims description 26
- 238000004458 analytical method Methods 0.000 claims description 26
- 230000004083 survival effect Effects 0.000 claims description 22
- 230000037361 pathway Effects 0.000 claims description 20
- 238000012549 training Methods 0.000 claims description 12
- 239000012530 fluid Substances 0.000 claims description 11
- 238000002075 inversion recovery Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 9
- 238000007405 data analysis Methods 0.000 claims description 8
- 238000000611 regression analysis Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 7
- 238000002271 resection Methods 0.000 claims description 7
- 238000004088 simulation Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000011119 multifactor regression analysis Methods 0.000 claims description 6
- 238000011088 calibration curve Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 3
- 238000011127 radiochemotherapy Methods 0.000 claims description 3
- 238000011282 treatment Methods 0.000 abstract description 23
- 238000010191 image analysis Methods 0.000 abstract description 2
- 208000020372 Infective dermatitis associated with HTLV-1 Diseases 0.000 description 19
- 101710177984 Isocitrate dehydrogenase [NADP] Proteins 0.000 description 19
- 101710102690 Isocitrate dehydrogenase [NADP] cytoplasmic Proteins 0.000 description 19
- 102100037845 Isocitrate dehydrogenase [NADP], mitochondrial Human genes 0.000 description 19
- 101710175291 Isocitrate dehydrogenase [NADP], mitochondrial Proteins 0.000 description 19
- 101710157228 Isoepoxydon dehydrogenase patN Proteins 0.000 description 19
- 238000002595 magnetic resonance imaging Methods 0.000 description 19
- 206010018338 Glioma Diseases 0.000 description 14
- 239000000090 biomarker Substances 0.000 description 13
- 238000002598 diffusion tensor imaging Methods 0.000 description 12
- 208000032612 Glial tumor Diseases 0.000 description 11
- 230000008236 biological pathway Effects 0.000 description 11
- 239000011159 matrix material Substances 0.000 description 11
- 230000002596 correlated effect Effects 0.000 description 8
- 238000003745 diagnosis Methods 0.000 description 7
- 230000006872 improvement Effects 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 201000010099 disease Diseases 0.000 description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000010201 enrichment analysis Methods 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 208000003174 Brain Neoplasms Diseases 0.000 description 4
- 102100025825 Methylated-DNA-protein-cysteine methyltransferase Human genes 0.000 description 4
- 238000003559 RNA-seq method Methods 0.000 description 4
- 238000002512 chemotherapy Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000014509 gene expression Effects 0.000 description 4
- 230000035772 mutation Effects 0.000 description 4
- 230000035755 proliferation Effects 0.000 description 4
- 238000001959 radiotherapy Methods 0.000 description 4
- 102000001301 EGF receptor Human genes 0.000 description 3
- 108060006698 EGF receptor Proteins 0.000 description 3
- BPEGJWRSRHCHSN-UHFFFAOYSA-N Temozolomide Chemical compound O=C1N(C)N=NC2=C(C(N)=O)N=CN21 BPEGJWRSRHCHSN-UHFFFAOYSA-N 0.000 description 3
- 238000013459 approach Methods 0.000 description 3
- 210000003169 central nervous system Anatomy 0.000 description 3
- 230000006378 damage Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000007717 exclusion Effects 0.000 description 3
- 239000000835 fiber Substances 0.000 description 3
- 239000011229 interlayer Substances 0.000 description 3
- 239000010410 layer Substances 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 108040008770 methylated-DNA-[protein]-cysteine S-methyltransferase activity proteins Proteins 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 229960004964 temozolomide Drugs 0.000 description 3
- 206010059866 Drug resistance Diseases 0.000 description 2
- 101150020771 IDH gene Proteins 0.000 description 2
- 102000012011 Isocitrate Dehydrogenase Human genes 0.000 description 2
- 108010075869 Isocitrate Dehydrogenase Proteins 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 201000007455 central nervous system cancer Diseases 0.000 description 2
- 208000025997 central nervous system neoplasm Diseases 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 210000000349 chromosome Anatomy 0.000 description 2
- 230000000875 corresponding effect Effects 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 230000034994 death Effects 0.000 description 2
- 238000002059 diagnostic imaging Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000007917 intracranial administration Methods 0.000 description 2
- 230000011987 methylation Effects 0.000 description 2
- 238000007069 methylation reaction Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000008506 pathogenesis Effects 0.000 description 2
- 230000002980 postoperative effect Effects 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 210000004872 soft tissue Anatomy 0.000 description 2
- 238000001356 surgical procedure Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000002560 therapeutic procedure Methods 0.000 description 2
- 210000001519 tissue Anatomy 0.000 description 2
- 210000004881 tumor cell Anatomy 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 210000004885 white matter Anatomy 0.000 description 2
- 101000796998 Bacillus subtilis (strain 168) Methylated-DNA-protein-cysteine methyltransferase, inducible Proteins 0.000 description 1
- 206010061445 Cerebral cyst Diseases 0.000 description 1
- 206010008111 Cerebral haemorrhage Diseases 0.000 description 1
- 206010009944 Colon cancer Diseases 0.000 description 1
- 208000001333 Colorectal Neoplasms Diseases 0.000 description 1
- 101150027068 DEGS1 gene Proteins 0.000 description 1
- 230000005778 DNA damage Effects 0.000 description 1
- 231100000277 DNA damage Toxicity 0.000 description 1
- 101710088194 Dehydrogenase Proteins 0.000 description 1
- 101000960235 Dictyostelium discoideum Isocitrate dehydrogenase [NADP] cytoplasmic Proteins 0.000 description 1
- 208000021994 Diffuse astrocytoma Diseases 0.000 description 1
- 206010064571 Gene mutation Diseases 0.000 description 1
- 101001082073 Homo sapiens Interferon-induced helicase C domain-containing protein 1 Proteins 0.000 description 1
- 208000014877 IDH-wildtype glioblastoma Diseases 0.000 description 1
- 102100027353 Interferon-induced helicase C domain-containing protein 1 Human genes 0.000 description 1
- 102000004310 Ion Channels Human genes 0.000 description 1
- 238000001337 Mantel test Methods 0.000 description 1
- 206010028851 Necrosis Diseases 0.000 description 1
- 206010061902 Pancreatic neoplasm Diseases 0.000 description 1
- 238000010220 Pearson correlation analysis Methods 0.000 description 1
- 206010039203 Road traffic accident Diseases 0.000 description 1
- XSTXAVWGXDQKEL-UHFFFAOYSA-N Trichloroethylene Chemical compound ClC=C(Cl)Cl XSTXAVWGXDQKEL-UHFFFAOYSA-N 0.000 description 1
- 208000036142 Viral infection Diseases 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000003556 assay Methods 0.000 description 1
- 230000008827 biological function Effects 0.000 description 1
- 230000031018 biological processes and functions Effects 0.000 description 1
- 210000005013 brain tissue Anatomy 0.000 description 1
- 238000007475 c-index Methods 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 230000003915 cell function Effects 0.000 description 1
- 206010008118 cerebral infarction Diseases 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000013211 curve analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000017188 evasion or tolerance of host immune response Effects 0.000 description 1
- 201000001169 fibrillary astrocytoma Diseases 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000010362 genome editing Methods 0.000 description 1
- 238000011331 genomic analysis Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000012165 high-throughput sequencing Methods 0.000 description 1
- 230000002962 histologic effect Effects 0.000 description 1
- 230000001506 immunosuppresive effect Effects 0.000 description 1
- 238000009169 immunotherapy Methods 0.000 description 1
- 230000008595 infiltration Effects 0.000 description 1
- 238000001764 infiltration Methods 0.000 description 1
- 230000002757 inflammatory effect Effects 0.000 description 1
- 239000003112 inhibitor Substances 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000005415 magnetization Effects 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 230000003211 malignant effect Effects 0.000 description 1
- 208000015486 malignant pancreatic neoplasm Diseases 0.000 description 1
- 230000017074 necrotic cell death Effects 0.000 description 1
- 210000004498 neuroglial cell Anatomy 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 201000002528 pancreatic cancer Diseases 0.000 description 1
- 208000008443 pancreatic carcinoma Diseases 0.000 description 1
- 239000012188 paraffin wax Substances 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 238000010837 poor prognosis Methods 0.000 description 1
- 238000011248 postoperative chemotherapy Methods 0.000 description 1
- 239000002243 precursor Substances 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007480 sanger sequencing Methods 0.000 description 1
- 150000003384 small molecules Chemical class 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000002626 targeted therapy Methods 0.000 description 1
- 230000008685 targeting Effects 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 210000001364 upper extremity Anatomy 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 230000009385 viral infection Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 239000001993 wax Substances 0.000 description 1
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
- G16H50/20—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
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B20/00—ICT specially adapted for functional genomics or proteomics, e.g. genotype-phenotype associations
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
- G16B30/00—ICT specially adapted for sequence analysis involving nucleotides or amino acids
-
- 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
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- 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
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/30096—Tumor; Lesion
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The application relates to the technical field of image analysis, in particular to a method and a system for risk stratification of GBM patients based on a conventional MRI sequence, wherein the method comprises the following steps: acquiring a data set of a GBM patient; image acquisition is carried out on the data set to obtain a conventional MRI sequence; based on the data set, acquiring transcriptome sequencing data; extracting image histology features based on conventional MRI sequences; constructing an image clinical model based on the image histology characteristics; obtaining an image histology analysis result based on the image clinical model and transcriptome sequencing data; and analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path, and explaining a biological basis behind the image histology characteristics. The application utilizes the conventional MRI sequence to perform risk stratification and prognosis prediction on IDH wild glioblastoma patients, and excavates biological passages behind the prognostic characteristics of image histology, thereby providing guidance for clinical accurate treatment.
Description
Technical Field
The embodiment of the application relates to the technical field of image analysis, in particular to a method and a system for risk stratification of GBM patients based on a conventional MRI sequence.
Background
Glioblastoma (GBM) is the most common intracranial malignancy with extremely high mortality, derived from glial cells or their precursors. The central brain tumor statistics department (The Central Brain Tumor Registry of the United States, CBTRUS) in 2022 issued 2015-2019 primary brain and other central nervous system tumor data, GBM was 50.1% of primary malignant brain tumors, median survival was 8 months, and 5-year survival was 6.9%. Current consensus treatments for GBM are the largest range of surgical resections of tumor mass followed by standard radiation therapy and Temozolomide (TMZ) chemotherapy. Although effective for prolonged survival through the positive treatments described above, the median survival time of GBM patients is still less than 15 months.
The poor prognosis of GBM is driven by a number of factors, the heterogeneity of tumor cells and the formation of immunosuppressive tumor microenvironment being the primary factors contributing to its malignant proliferation and resistance development. On the one hand, the complex heterogeneity of GBM cells is reflected in complex mutations of genes and disturbances of biological pathways, promoted by the local inflammatory tumor microenvironment, which mainly induces invasiveness and drug resistance of the tumor; on the other hand, the immune tumor microenvironment of GBM provides multiple ways for the immune escape of tumor cells, which together lead to the proliferation and drug resistance increase of GBM cells, and the progress of the disease is accelerated. In recent years, emerging therapies for GBM such as: targeting therapies, immunotherapy, gene editing, small molecule inhibitors, etc., have been on the go, but these therapeutic improvements and exploration have not led to shortcuts for a significant increase in patient survival.
Effective GBM biomarkers are determined and screened, the treatment targets of GBM are explored, the pathogenesis of the GBM is clarified, and the GBM biomarker is important to the improvement of the survival period of GBM patients. The magnetic resonance imaging (Magnetic Resonance Imaging, MRI) technique is widely used for diagnosing various clinical system diseases due to its advantages such as non-invasiveness, good soft tissue resolution, and no radiation damage. MRI is recommended as the first examination in the diagnosis of GBM and as the primary follow-up examination after surgery. Today, the prior art mostly employs multi-modal advanced MRI sequences, such as: dynamic magnetization rate contrast (DSC), diffusion Tensor Imaging (DTI), and Magnetic Resonance Spectroscopy (MRS), multi-modal advanced MRI sequences have been demonstrated to exert the advantages of different modalities in diagnosis and treatment of GBM, improving GBM diagnostic rate; however, conventional MRI sequences remain the most widely used tumor screening sequences in the clinic. Therefore, developing an image biomarker based on the MRI routine sequence to perform early noninvasive risk stratification on the wild type GBM of isocitrate dehydrogenase (Tsocitrate Dehydrogenase, IDH) has important practical significance and popularization value.
Disclosure of Invention
The embodiment of the application provides a method and a system for risk stratification of GBM patients based on a conventional MRI sequence, which utilize the image histology characteristics of the conventional MRI sequence to carry out risk stratification and prognosis prediction on IDH wild type glioblastoma patients, and excavate biological passages behind the image histology prognosis characteristics, thereby providing important guidance for clinical accurate treatment.
To solve the above technical problems, in a first aspect, an embodiment of the present application provides a method for risk stratification of GBM patients based on a conventional MRI sequence, comprising the steps of: firstly, acquiring a data set of a GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set; then, carrying out image acquisition on the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence; next, based on the conventional MRI sequence, extracting image histology features; constructing an image clinical model based on the image histology characteristics; then, based on the image clinical model and the transcriptome sequencing data, obtaining an image histology analysis result; next, analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; finally, the biological basis behind the image group biological characteristics is explained based on the image clinical model and the cross passage.
In some exemplary embodiments, image histology features are extracted based on the conventional MRI sequence; and constructing an image clinical model based on the image histology features, comprising: preprocessing the conventional MRI sequence to obtain a preprocessed conventional MRI sequence; extracting image histology features from each three-dimensional volume of interest based on the preprocessed conventional MRI sequences; the three-dimensional interesting volume is a three-dimensional volume for generating tumors through three-dimensional simulation; the image histology features comprise texture histology features, shape histology features and intensity histology features; screening out the optimal image histology characteristics from the data set based on the image histology characteristics to form an image histology tag; and constructing an image clinical model based on the image histology tag.
In some exemplary embodiments, the image histology signature is used to predict prognosis of survival for IDH wild-type GBM patients.
In some exemplary embodiments, selecting the best image histology feature in the dataset based on the image histology feature to form an image histology tag comprises: calculating intra-group correlation coefficients of each image group learning feature, and screening out features with sketch consistency larger than a threshold value based on the intra-group correlation coefficients to obtain primary screening features; carrying out single factor regression analysis on the primary screening characteristics, and counting the consistency index of the primary screening characteristics to obtain secondary screening characteristics; and screening the secondary screening characteristics again by adopting a minimum absolute shrinkage and selection operator screening method, and screening out the optimal image histology characteristics to form an image histology label.
In some exemplary embodiments, constructing an imaging clinical model based on the imaging histology features includes: constructing an image group learning risk score based on the image group learning features; incorporating clinical factors into a multi-factor regression analysis model, and establishing a clinical model for predicting total survival; the clinical factors include sex, age, surgical resection degree, whether chemoradiotherapy and preoperative Carlsberg score; and adding the image group academic risk score into the clinical factors to obtain new clinical factors, and incorporating the new clinical factors into the multi-factor regression analysis model to construct an image clinical model.
In some exemplary embodiments, after constructing an imaging clinical model based on the imaging histology features, before deriving imaging histology analysis results based on the imaging clinical model and the transcriptome sequencing data, further comprising: and evaluating the image clinical model.
In some exemplary embodiments, evaluating the imaging clinical model includes: and evaluating the image clinical model by adopting a survival analysis curve, a calibration curve and a decision curve respectively.
In some exemplary embodiments, interpreting the biological basis behind the image set of features based on the image clinical model and the intersection path includes: based on the image clinical model and the cross path, the biological basis under different risk stratification, the biological meaning of the single image group biological feature and the biological meaning of the single MRI sequence are respectively revealed to explain the biological basis behind the image group biological feature.
In a second aspect, embodiments of the present application also provide a system for risk stratification of GBM patients based on conventional MRI sequences, comprising: the system comprises a data set module, a conventional MRI sequence acquisition module, a model construction module and a data analysis module which are connected in sequence; the data set module is used for acquiring a data set of the GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set; the conventional MRI sequence acquisition module is used for acquiring images of the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence; the model building module is used for extracting image histology characteristics according to the conventional MRI sequence; constructing an image clinical model based on the image histology characteristics; the data analysis module is used for obtaining an image histology analysis result according to the image clinical model and the transcriptome sequencing data; analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; and explaining the biological basis behind the image group biological characteristics based on the image clinical model and the cross passage.
In some exemplary embodiments, the model building module includes a preprocessing module, a feature extraction module, an image histology labeling module, and an image clinical model module connected in sequence; the pretreatment module is used for carrying out pretreatment on the conventional MRI sequence to obtain a pretreated conventional MRI sequence; the feature extraction module is used for extracting image histology features from each three-dimensional interested volume according to the preprocessed conventional MRI sequences; the three-dimensional interesting volume is a three-dimensional volume for generating tumors through three-dimensional simulation; the image histology features comprise texture histology features, shape histology features and intensity histology features; the image histology tag module is used for screening out the optimal image histology features from the data set according to the image histology features so as to form an image histology tag; and the image clinical model module is used for constructing an image clinical model according to the image histology tag.
The technical scheme provided by the embodiment of the application has at least the following advantages:
the embodiment of the application provides a method and a system for risk stratification of GBM patients based on a conventional MRI sequence, wherein the method comprises the following steps: firstly, acquiring a data set of a GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set; then, carrying out image acquisition on the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence; next, based on the conventional MRI sequence, extracting image histology features; constructing an image clinical model based on the image histology characteristics; then, based on the image clinical model and the transcriptome sequencing data, obtaining an image histology analysis result; next, analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; finally, the biological basis behind the image group biological characteristics is explained based on the image clinical model and the cross passage.
According to the method for risk stratification of the GBM patient based on the conventional MRI sequence, the risk stratification and prognosis prediction are carried out on the IDH wild Glioblastoma (GBM) patient by utilizing the image histology characteristics from the conventional MRI sequence, the biological meaning of the image histology characteristics is explored by adopting a reliable image genomics analysis technology, and the biological path behind the image histology prognosis characteristics is excavated, so that important guidance is provided for clinical accurate treatment.
Drawings
One or more embodiments are illustrated by way of example and not limitation in the figures of the accompanying drawings, which are not to be construed as limiting the embodiments unless specifically indicated otherwise.
FIG. 1 is a flow chart of a method for risk stratification of GBM patients based on a conventional MRI sequence according to an embodiment of the present application;
FIG. 2A is a diagram illustrating data acquisition according to an embodiment of the present application;
FIG. 2B is a schematic diagram illustrating image histology feature screening and analysis according to an embodiment of the present application;
FIG. 2C is a schematic illustration of GSEA analysis according to an embodiment of the application;
FIG. 2D is a schematic diagram of a WGCNA assay according to an embodiment of the application;
FIG. 2E is a schematic diagram of a cross-over pathway and individual feature biological disclosure, sequence biological disclosure provided by an embodiment of the present application;
fig. 3 is a schematic diagram of a system for risk stratification of GBM patients based on conventional MRI sequences according to an embodiment of the present application.
Detailed Description
From the background, the prior art often adopts a multi-mode advanced MRI sequence in GBM diagnosis and treatment, neglects the function and the value of the conventional MRI sequence, and lacks the prognostic value and biological meaning of systematically researching the high-flux image histology characteristics derived from the conventional MRI sequence.
At present, diagnosis and treatment improvement and exploration of GBM do not bring shortcuts to the remarkable increase of survival rate of patients, so that effective GBM biomarkers are determined and screened, the treatment targets of GBM are explored, the pathogenesis of the GBM is clarified, and the diagnosis and exploration are of great importance to the improvement of the survival period of GBM patients.
Currently, biomarkers known to be significantly correlated with GBM prognosis are O6-methylguanine-DNA methyltransferase (MGMT), epidermal Growth Factor Receptor (EGFR), isocitrate dehydrogenase 1/2 (IDH 1/2) and short arm of chromosome 1/long arm of chromosome 19 (1 p/19 q). Studies have shown that patients with higher levels of MGMT methylation have more pronounced effects on temozolomide in chemotherapy than patients with no or lower levels of MGMT methylation. EGFR amplification is closely related to the development of GBM, and EGFR-targeted therapies have seen initial success in colorectal and pancreatic cancer in recent years, but have not been satisfactory in clinical trials of GBM. More recently, IDH mutations and 1p/19q co-deletion status have been considered as key molecular diagnostic markers in the adult diffuse glioma classification in the latest 2021 world health organization central nervous system tumor 5 th edition classification WHO CNS 5). According to GBM diagnostic criteria for 2021WHO CNS5, IDH mutant GBMs previously classified by version 2016 WHO CNS4 were excluded from the class of GBMs, and new criteria incorporated a part of diffuse astrocytomas previously identified as IDH wild-type. This recent classification underscores the critical role of biomarkers in GBM diagnostics. However, the availability of these biomarkers mainly depends on invasive procedures and expensive detection means, which results in a certain degree of inability to be widely developed in clinical work, and thus, some patients cannot get accurate diagnosis and treatment in time, and there is a strong need for an alternative biomarker. In addition, recent studies have found that IDH wild-type GBM patients have clinical outcome of heterogeneity with potential risk stratification in the interior. Therefore, the development of a noninvasive and efficient biological marker for predicting and risk stratification of the WHO CNS 5-based IDH wild-type GBM is of great practical significance for the personalized treatment and management of future tumor patients.
MRI is one of the most widely used medical imaging tools in clinic at present, has the advantages of non-invasiveness, no radiation damage, high resolution, multiple parameters and high contrast ratio of soft tissues, and is also the most commonly used imaging means for diagnosing glioma. As a structural imaging sequence in MRI, diffusion tensor imaging DTI is the only advanced sequence capable of displaying brain fiber bundle structure in living body in MRI at present, and utilizes anisotropy of water molecular diffusion motion in brain tissue to display microstructure characteristics of brain white matter fiber bundles, reflecting the damage and infiltration degree of brain white matter fiber bundles by tumor. DTI contains four main dispersion indicators: average diffusivity (MD), anisotropy fraction (Fractional Anisotropy, FA), axial diffusivity (Axial Diffusivity, AD), and radial diffusivity (Radial Diffusivity, RD), which have previously been demonstrated to be useful in predicting glioma classification, judging treatment response, and assessing patient prognosis. However, these studies only use semi-quantitative DTI indices for histogram analysis to predict patient prognosis, lacking systematic studies of prognostic value and biological implications of high-throughput imaging histology features derived from DTI sequences.
In recent years, with the improvement of medical imaging technology and the level of large-scale high-throughput sequencing technology, the emerging interdisciplinary of image genomics is becoming a rapidly growing research field. The method combines high-flux data in medical images with disease-related gene data to perform comprehensive analysis and mining, and aims to explore potential relations between medical image data and disease molecular characteristics. It makes a prominent contribution in explaining the biological basis of the image histology. Imaging genomic studies of past gliomas have shown that the imaging prognostic features derived from magnetic resonance conventional sequences are associated with specific biological pathways. These studies typically use only a single differential gene enrichment analysis method to find the biological implications of the image histology features.
Image histology (Radiomics) is a technique that converts visual medical images into an array matrix of subvision to quantify the relevant phenotypes. The imaging group science can be used for converting visual medical images into steady sub-visual digital indexes, so that massive potential relations between medical images and tumors are excavated, and a new thought is provided for prognosis prediction of tumors and molecular target spot exploration. Through recent development, GBM noninvasive biomarkers constructed by using imaging histology have been demonstrated to be able to effectively conduct risk stratification guidance prognosis. In summary, image histology has great potential in fully mining tumor phenotype information.
Image genomics (Radiogenomics) is a rapidly evolving field based on genomics, which combines image and genomics, providing biological interpretability for data-driven features of image genomics. The imaging genomics establishes the connection between the imaging characteristics with prognostic significance and molecular biological pathways and genes, is widely applied to the tumor field, and has important values in the aspects of predicting classification of tumors, molecular typing, treatment target discovery and guiding individuation prognosis. Recently, several studies have revealed the relationship between the imaging histology characteristics and the biological basis of GBM using imaging genomics techniques.
Although image genomics based on genetic and molecular properties opens up a biological interpretable "black box" of high-throughput histologic data, the stability and reliability of its clinical popularization still deserves further verification. Previous imaging genomics studies have revealed biological pathways behind imaging genomics features by Gene set enrichment analysis (Gene-set enrichment analysis, GSEA) or weighted Gene co-expression network analysis (Weighted correlation network analysis, WGCNA) methods, respectively. Even though both methods consider all genes in the queue, their enrichment direction is of great importance. The goal of GSEA is to determine if genes in a gene set tend to be enriched at the top or bottom of a pre-defined list of genes, whereas WGCNA focuses on finding a set (module) of genes that are co-expressed (have a consistent trend of variation) throughout the genes. The related art proposes that in MDA5 DM cases, the same immune-related process is determined using both GSEA and WGCNA methods, enhancing the robustness of the enriched pathway. However, few studies have used both GSEA and WGCNA methods in GBM imaging genomics to study the biological pathways behind the imaging genomic phenotype. Therefore, the combined analysis method of GSEA and WGCNA is adopted in the image genomics analysis, so that the defects of the prior research can be overcome, and the robustness of biological interpretation after the image genomics characteristics is improved.
The imaging histology features from a single conventional MRI sequence show good prognostic value in gliomas, but are rarely investigated based on the biological significance behind a single sequence. From an image anatomical point of view, different conventional MRI sequences are associated with the dominant tumor imaging morphology of gliomas, but there is currently insufficient biological evidence to support this view.
In summary, the prior art has the following drawbacks: on the one hand, the prior art relies solely on the imaging histology characteristics of conventional MRI sequences to risk stratify GBM, while ignoring the role and value of DTI sequences. In addition, prior art studies on DTI generally only use semi-quantitative DTI indices for histogram analysis to predict patient prognosis, lacking systematic studies on prognostic value and biological implications of high-throughput imaging histology features derived from DTI sequences. Additionally, the prior art generally uses only a single differential gene enrichment analysis method to find the biological meaning of an image histology signature when studying the biological interpretation of the image histology prognosis signature. However, differential gene enrichment analysis is more focused on genes with significant differential expression, ignoring genes with insignificant differential expression but significant biological significance. In addition, the number of genes is numerous and complex, and the cooperative relationship between genes is a point that must be considered in genomics analysis, which is ignored by the prior art. Finally, the prior art research on GBM is mostly limited to the WHO glioma classification approach of version 2016, but does not employ the WHO glioma classification approach of version 2021.
In order to solve the above technical problems, an embodiment of the present application provides a method and a system for risk stratification for GBM patients based on a conventional MRI sequence, the method comprising: firstly, acquiring a data set of a GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set; then, carrying out image acquisition on the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence; next, based on the conventional MRI sequence, extracting image histology features; constructing an image clinical model based on the image histology characteristics; then, based on the image clinical model and the transcriptome sequencing data, obtaining an image histology analysis result; next, analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; finally, the biological basis behind the image group biological characteristics is explained based on the image clinical model and the cross passage. The application provides a method and a system for risk stratification of GBM patients based on a conventional MRI sequence, which utilize the image histology characteristics of the conventional MRI sequence from magnetic resonance imaging to carry out risk stratification and prognosis prediction on IDH wild glioblastoma patients, and excavate biological passages behind the image histology prognosis characteristics, thereby providing important guidance for clinical accurate treatment.
Embodiments of the present application will be described in detail below with reference to the attached drawings. However, it will be understood by those of ordinary skill in the art that in various embodiments of the present application, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, the claimed technical solution of the present application can be realized without these technical details and various changes and modifications based on the following embodiments.
Referring to fig. 1, an embodiment of the present application provides a method for risk stratification of GBM patients based on conventional MRI sequences, comprising the steps of:
step S1, acquiring a data set of a GBM patient; GBM patients are glioblastoma patients; the data set includes a training set and a testing set.
S2, performing image acquisition on the data set to obtain a conventional MRI sequence; based on the data set, acquiring transcriptome sequencing data; conventional MRI sequences include fluid attenuation inversion recovery imaging sequences, T1 weighted contrast enhancement imaging sequences, T1 weighted imaging sequences, and T2 weighted imaging sequences.
S3, extracting image histology characteristics based on a conventional MRI sequence; and constructing an image clinical model based on the image histology characteristics.
And step S4, obtaining an image histology analysis result based on the image clinical model and the transcriptome sequencing data.
And S5, analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross passage for revealing biological meanings of the image histology characteristics.
Step S6, explaining the biological basis behind the image group biological characteristics based on the image clinical model and the cross passage
Because the prior art mostly adopts a multi-mode advanced MRI sequence in diagnosis and treatment of GBM, the application and the value of a conventional MRI sequence are ignored, and the application aims at the technical problem, and provides a method for risk stratification of GBM patients based on the conventional MRI sequence, wherein the conventional MRI sequence comprises fluid attenuation inversion recovery imaging (FLAIR), T1 weighted contrast enhancement imaging (T1 c), T1 weighted imaging (T1) and T2 weighted imaging (T2). The application provides important guidance for clinical accurate treatment by using the image histology characteristics from the conventional MRI sequences (T1, T2, T1c and FLAIR) to perform risk stratification and prognosis prediction on IDH wild Glioblastoma (GBM) patients and excavating biological pathways behind the image histology prognosis characteristics.
The application aims to develop a noninvasive biological marker for predicting prognosis of IDH wild type GBM based on MRI conventional sequences, and explores biological meaning of image histology characteristics by adopting reliable image genomics analysis technology. Furthermore, the present application is based on the latest WHO glioma classification method of 2021 edition to study the heterogeneity of IDH wild-type GBM.
FIGS. 2A-2E are flow charts illustrating stages of a method for risk stratification of GBM patients based on conventional MRI sequences, respectively, provided by the present application; the method provided by the application mainly comprises the following five parts: A. collecting data; B. constructing and verifying an image histology model; C. image genomics GSEA analysis; D. image genomics WGCNA analysis; E. the prognostic phenotype biological basis is revealed. The data acquisition process is a process of acquiring a data set of the GBM patient, then carrying out image acquisition on the data set to obtain a conventional MRI sequence (T1, T2, T1c and FLAIR), and extracting image histology characteristics from the conventional MRI sequence; the present application developed and validated an image histology prognostic model based on LASSO algorithm (Least Absolute Shrinkage and Selection Operator, LASSO) to predict prognosis of IDH wild-type GBM patients and explored the explanation of the biological basis behind image histology features using both GSEA and WGCNA methods. Compared with the prior art, the method for risk stratification of GBM patients based on the conventional MRI sequences has great advantages. In one aspect, the application exploits the prognostic value and biological implications of high-throughput imaging histology features derived from conventional MRI sequences; in another aspect, the present application exploits the biological implications of image genomic features using reliable image genomic analysis techniques. Furthermore, the present application is a study of the heterogeneity of IDH wild-type GBM according to the latest WHO glioma classification approach of 2021 edition.
In step S1, firstly, a data set of GBM patients is acquired; GBM patients are glioblastoma patients; the data set includes a training set and a testing set. Specifically, the data set of GBM patients is acquired by setting "patient cohort selection criteria". The main cohort of the study was a case where brain tumors were surgically resected in hospital neurosurgery and diagnosed as IDH wild-type GBM according to CNS 5. Wherein the patient cohort selection criteria include a case inclusion criterion and a case exclusion criterion, the case inclusion criterion and the case exclusion criterion being respectively as follows:
case inclusion criteria were as follows:
(1) Adult patients (age. Gtoreq.18 years).
(2) Preoperative conventional MRI image data, including T1, T2, T1c and FLAIR.
(4) The time for patient to undergo standardized surgical resection is within 14 days after imaging.
(5) The clear IDH wild type status and histopathological results, or the presence of Formalin-fixed wax block embedded (Formalin-Fixed Paraffin Embedded, FFPE) tissue for subsequent detection to further clear the tumor IDH mutation status and vascular proliferation or tissue necrosis.
(6) No glioma-related treatments (e.g., chemotherapy, radiation therapy, etc.) were performed prior to MRI.
Case exclusion criteria were as follows:
(1) MRI images are not acquired on a high field magnetic resonance scanner (Siemens 3.0T) (e.g., philips 3.0T or 1.5T).
(2) Patients are accompanied by other intracranial diseases (such as multiple cerebral infarction, cerebral hemorrhage, and cerebral cyst).
(3) Preoperative MRI images are of poor quality (e.g., have significant motion artifacts, etc.).
(4) Patient basic clinical information is missing (e.g., lost visit, etc.).
(5) Patient death due to non-GBM causes (e.g., surgical accidents, traffic accidents, etc.).
The basic clinical information content is as follows:
age, sex, pre-operative karst score (Karnofsky Performance Score, KPS), tumor resection, post-operative radiotherapy condition, post-operative chemotherapy condition, survival status, total Survival (OS). ( And (3) injection: the total survival time is calculated in months with the date of surgery as the time start of the follow-up, the date of death as the end of the follow-up. )
In some embodiments, molecular information and transcriptome RNA sequencing operations are performed after the GBM patient dataset is obtained. The molecular information is from two sources, one is that part of patients check the mutation state of IDH gene in postoperative pathology detection, and the molecular information of the part of patients comes from the hospital electronic file system. Another portion of patients did not undergo this examination, and FFPE samples from the corresponding patients were borrowed from the pathologist by members of the team, and IDH gene mutation status was detected using Sanger sequencing. In addition, fresh frozen tumor specimens from several patients (n=132) of the group were randomly selected for transcriptome Sequencing (RNA Sequencing, RNA-Seq) to obtain transcriptome Sequencing data of the tumor specimens.
In some embodiments, the image acquisition is performed on the dataset in step S2, resulting in a conventional MRI sequence; wherein an image acquisition is performed using a magnetic resonance (Magnetic Resonance, MR) device. All patients in the group were image acquired in a hospital magnetic resonance complex operating room by intra-operative high field strength 3.0T MR scanner (instrument model: siemens Healthcare, magnethom Verio, erlangen) using 12 channel coils matched to the instrument. Before entering the examination room, the patient routinely removes all metallic products on his body, avoiding metallic artifacts on the images during machine operation and causing major operating room safety accidents. In the image acquisition process, the head of the patient is fixed at the center of the coil, and the patient is ordered to avoid the head and the upper limb from moving in the whole course, so that the motion artifact in the image is reduced.
The magnetic resonance imaging protocol includes T1WI, T2WI, CE-TIWI, FLAIR and DTI, the specific parameters are as follows:
the parameters of T1WI and CE-TIWI are 165 ms-280 ms of time of cycle (TR); the echo Time of Echo (TE) is 2.46 ms-4.76 ms; the scan Field (FOV) was 220X 172mm 2 ～240×240mm 2 The method comprises the steps of carrying out a first treatment on the surface of the The thickness of the section layer is 5mm; the interlayer spacing was 1mm.
The T2WI parameters are: TR is 3600 ms-6300 ms; TE is 82 ms-125 ms; FOV is 220X 162mm 2 ～240×240mm 2 The method comprises the steps of carrying out a first treatment on the surface of the The thickness of the section layer is 5mm; the interlayer spacing was 1mm.
The FLAIR parameters were: TR is 6000ms; TE is 94ms; FOV is 220X 193mm 2 The method comprises the steps of carrying out a first treatment on the surface of the The thickness of the section layer is 5mm; the interlayer spacing was 1mm.
In some embodiments, constructing an imaging clinical model based on the conventional MRI sequence in step S3 includes:
step S301, preprocessing the conventional MRI sequence to obtain a preprocessed conventional MRI sequence.
Step S302, extracting image histology characteristics from each three-dimensional interested volume based on the preprocessed conventional MRI sequence; the three-dimensional interesting volume is a three-dimensional volume for generating tumors through three-dimensional simulation; the image histology features include texture histology features, shape histology features, and intensity histology features.
Step S303, based on the image histology characteristics, the best image histology characteristics are screened out from the training set to form an image histology label.
And step 304, constructing an image clinical model based on the image histology tag.
Specifically, in step S301, the acquired MRI data is first subjected to standardized preprocessing. The preprocessing starts with offset field distortion correction for N4 ITK. Secondly, the first step of the method comprises the steps of, all voxels are interpolated by tri-line interpolation equidirectional resampling is 1X 1mm 3 Is included in the set of (a) voxels. Conventional MR images of each patient are rigidly registered in 3D SliceAnd r software, using the T1c of axial resampling as a template, and adopting mutual information similarity indexes to generate registration images which are respectively marked as rFLAIR, rT1c, rT1 and rT2. Finally, histogram matching is used to normalize the gray scale distribution.
Note that, rvair, rT1c, rT1, and rT2 are registered images of fluid attenuation inversion recovery imaging (FLAIR), T1 weighted contrast enhancement imaging (T1 c), T1 weighted imaging (T1), and T2 weighted imaging (T2), respectively.
Step S302 is mainly a process of extracting the image histology features. Wherein the three-dimensional volume of interest (Volume of Interest, VOI) is a three-dimensional volume that generates a tumor by three-dimensional simulation; the image histology features include texture histology features, shape histology features, and intensity histology features. Specifically, the PyCharm software was used to extract image histology features in a Python environment (version: 3.7.0) using the Pyradio kits (version: 3.0.1). Extracting high-flux image histology features, namely image histology features, from each VOI respectively by means of a public database; the image histology features comprise texture histology features, shape histology features and intensity histology features. The texture features are of five types: (1) Gray Level Co-occurrence Matrix, GLCM; (2) A gray run matrix (Gray Level Run Length Matrix, GLRLM); (3) A gray area size matrix (Gray Level Size Zone Matrix, GLSZM); (4) A gray scale correlation matrix (Gray Level Dependence Matrix, GLDM); (5) Adjacent gray differential matrices (Neighborhood Gray Tone Difference Matrix, NGTDM). Meanwhile, in order to fully mine biological information of the MRI image as much as possible, two types of filters are used for obtaining texture group characteristics and intensity group characteristics from the original image and the converted image. Mainly comprises the following steps: (1) wavelet transformation; (2) Gauss-Laplace operator conversion, the operator contains four levels: 2.0, 3.0, 4.0 and 5.0. It should be noted that, the acquisition, processing, delineation and feature extraction of MRI images in the present application all follow the image biomarker standardization initiative (Image Biomarker Standardisation Initiative, IBSI) guidelines.
Step S303 is mainly a process of screening the image histology features to obtain the best image histology features to form the image histology tag. In some embodiments, the image histology signature is used to predict prognosis of survival for IDH wild-type GBM patients.
In some embodiments, screening the best image histology features in the dataset based on the image histology features in step S303 to form an image histology tag includes:
step S3031, calculating the intra-group correlation coefficient of each image group chemical characteristic, and screening out the characteristic with the sketching consistency larger than a threshold value based on the intra-group correlation coefficient to obtain a primary screening characteristic.
And step S3032, carrying out single factor regression analysis on the primary screening characteristics, and counting the consistency index of the primary screening characteristics to obtain secondary screening characteristics.
And step S3033, adopting a minimum absolute shrinkage and selection operator screening method to screen the secondary screening characteristics again, and screening out the optimal image histology characteristics so as to form an image histology label.
Specifically, when the image histology features are screened, features with higher consistence are screened and sketched first. According to the data set between 100 raters and the data set of the raters, selecting the characteristics with high stability and consistency, wherein ICC smaller than 0.4 is generally considered to be poor in consistency of VOI sketch, ICC larger than 0.75 is considered to be good in consistency of sketch, and the characteristics that ICC is larger than or equal to 0.85 are reserved, so that the screening characteristics are obtained for the next analysis.
Next, features highly correlated with survival prognosis are screened. The application carries out single factor Cox regression analysis on each feature in the primary screening features, and counts the consistency index (Concordance index, C-index); a C-index greater than 0.5 is set as a feature positively correlated with prognosis, and a C-index less than 0.5 is set as a feature negatively correlated with prognosis. The application reserves the characteristic that C-index is more than or equal to 0.55 (positive correlation) or index is less than or equal to 0.45 (negative correlation) (P is less than 0.05), and the secondary screening characteristic is obtained.
Then, the minimum absolute shrinkage and selection operator (Least absolute shrinkage and selection operator, LASSO) is used for screening again. And performing further dimension reduction treatment on the secondary screening features by using LASSO punishment Cox proportional risk regression, and selecting features with Lasso coefficients not being 0 to form optimal image histology features so as to form image histology tag subsequent input modeling.
In some embodiments, constructing an imaging clinical model based on the imaging histology features in step S3 includes: constructing an image group learning risk score based on the image group learning features; incorporating clinical factors into a multi-factor regression analysis model, and establishing a clinical model for predicting total survival; the clinical factors include sex, age, surgical resection degree, whether chemoradiotherapy and preoperative Carlsberg score; and adding the image group academic risk score into the clinical factors to obtain new clinical factors, and incorporating the new clinical factors into the multi-factor regression analysis model to construct an image clinical model.
Specifically, an image group risk score (Radscore) is first constructed: and constructing Radscore by adopting an LASSO-Cox regression model, selecting the optimal lambda value by using 10 times of cross validation (10-fold Cross Validation), determining the optimal regression model, extracting regression coefficients by a coef function, and obtaining a linear combination obtained by adding the sum of the products of the prognosis characteristic values and the corresponding coefficients of LASSO screening, namely Radscore. A multi-factor Cox regression analysis Clinical Model (CM) was constructed for the clinical information, while a multi-factor Cox regression analysis image-clinical model (R-CM) was constructed by incorporating Radscore into the clinical information model. Nomograms (nomograms) are used to show the predicted extent of contribution of each variable in the model to survival, respectively.
Next, CM is established: the gender, age, surgical resection degree, whether radiotherapy and chemotherapy and pre-operative KPS (Karnofsky performance status) scores among clinical factors (clinical information) are included in a multi-factor Cox regression analysis model, and a clinical model for predicting the OS is established.
Finally, R-CM is established: adding Radscore into clinical factors for establishing CM to obtain new clinical factors, and incorporating the new clinical factors into a multi-factor Cox regression analysis model to establish an image clinical model.
In some embodiments, after constructing an image clinical model based on the image histology features in step S3, before obtaining an image histology analysis result based on the image clinical model and the transcriptome sequencing data in step S4, the method further comprises: the image clinical model is evaluated.
In some embodiments, evaluating the imaging clinical model includes: and evaluating the image clinical model by adopting a survival analysis curve, a calibration curve and a decision curve respectively. Specifically, a Radscore-based critical (cutoff) value is obtained in the cine training set and then applied to the internal and external validation sets. The ability of Radscore to distinguish between high and low risk groups was assessed using a Kaplan Meier, K-M, curve in an internal and external validation set, respectively. The predictive performance of each model is measured by calculating C-index. A calibration curve (Calibration curve) is used to evaluate the consistency of predictions and observations in the alignment. Decision curves (Decision Curve Analysis, DCA) are used to measure clinical effectiveness of the model. The net weight classification improvement (Net Reclassification Improvement, NRI) and the red pool information criterion (Akaike Information Criterion, AIC) evaluate improved performance and potential risk of overfitting in the model, respectively.
In some embodiments, in step S5, two analysis methods, WGCNA and GSEA are used for weighted gene co-expression network analysis and gene set enrichment analysis, and transcriptome sequencing data is analyzed to obtain crossover pathways for revealing biological meaning of image histology features.
Specifically, when the GSEA analysis method is adopted to analyze data, the optimal cutoff value is utilized to divide the image genomics analysis set into high and low risk subgroups. Then, using R package "DESeq2" to identify the differentially expressed genes (Differentially expressed genes, DEGs) between the high and low risk subgroups and obtain a Fold difference (Fold change) of the sample gene expression level, log2 transforming and arranging in reverse order to obtain the ordered gene log 2 Fold Change data table. Next, a pathway enrichment analysis was performed based on Kyoto Encyclopedia of Genes and Genomes (KEGG), hallmark, reactome, bioCarta, pathway Interaction Database (PID) and WikiPathways (WP) gene sets in the MSigDB database (http:// www.gsea-MSigDB. Org/gsea/MSigDB /), with a false discovery rate (False discover rate, FDR) of less than0.05 was considered significantly enriched. Genomic variation analysis (Gene set variation analysis, GSVA) was used to quantify the activity of each enriched pathway to calculate the specific GSVA score for each pathway in each patient. A pearson correlation (Pearson correlation) was used to evaluate the correlation of each path GSVA value to Radscore, and paths with an error discovery rate FDR less than 0.01 were considered significantly correlated to Radscore.
Specifically, when data are analyzed using the WGCNA analysis method, first, WGCNA analysis is performed on the image genomics analysis set RNA-seq data, aggregating highly interconnected genes into several Gene modules (Gene modules) that may be involved in a common biological process. Then, the GSVA value of each module is calculated for that set of genes and pearson correlated with Radscore, where modules with a false discovery rate FDR less than 0.01 are considered significantly correlated with Radscore. For each Radscore-associated module, a pathway with a false discovery rate FDR of less than 0.01 was considered a pathway significantly associated with Radscore using an R-packet "clusterif iotaler" for pathway enrichment analysis in the KEGG, hallmark, reactome, bioCarta, PID and WP gene sets.
In some embodiments, the interpreting the biological basis behind the image histology features in step S6 based on the image clinical model and the intersection path includes: based on the image clinical model and the cross path, the biological basis under different risk stratification, the biological meaning of the single image group biological feature and the biological meaning of the single MRI sequence are respectively revealed to explain the biological basis behind the image group biological feature.
Specifically, the pathways significantly related to Radscore obtained by the two analysis methods of GSEA and WGCNA are intersected to be used as a robust pathway for revealing the biological significance of the prognostic phenotype, and each intersection pathway is divided into four major classes, namely immune-related, DNA damage-related, proliferation-related and other (viral infection, ion channel transport, transmitter transport and complex cell functions) one by one according to the biological functions of the intersection pathway. Wherein, different intersection paths respectively reveal the biological basis under different risk stratification, the biological meaning of a single image group biological feature and the biological meaning of a single MRI sequence, and the specific steps are as follows:
(1) Revealing the biological basis under different risk stratification: the expression activity of each intersection pathway in the high and low risk groups of the image genomics was quantified using GSVA, revealing the biological pathway differences between the high and low risk groups.
(2) Biological meaning of individual image histology features is revealed: inter-matrix pearson correlation analysis of prognostic imaging panel eigenvalues and intersection pathway GSVA values, which are involved in constructing Radscore, wherein FDR less than 0.05 indicates that the pathway is significantly correlated with the feature, and the potential links between biological pathways and individual features are analyzed in detail.
(3) The biological meaning of a single MRI sequence is revealed: the prognosis features participating in the formation of Radscore are grouped according to the belonging MRI sequences and inter-matrix Mantel (Mantel) test is performed with the GSVA values of the intersection path, wherein the Mantel test is a correlation analysis of the two matrices. Pathways with P less than 0.05 were considered to be significantly related to sequence, further revealing the potential association between biological pathways and MRI sequences.
Based on this, the present application establishes an imaging set of academic risk scores Radscore for noninvasively predicting and stratifying prognosis of adult IDH wild-type GBM patients based on preoperative MRI routine sequences (FLAIR, T1c, T1 and T2) for the IDH wild-type GBM class of the latest WHO CNS5 glioma classification. In addition, the Radscore robustness and universality are verified by an internal independent verification set and an external independent verification set, and the interpretability is provided for clinical application popularization. Finally, the present application reveals the correlation between different risk stratification and prognostic imaging histology phenotypes and biological pathways using intersection pathways obtained by two methods of GSEA and WGCNA in imaging genomics.
Compared with the prior art, the method for risk stratification of GBM patients based on the conventional MRI sequences has the advantages that: firstly, the application uses a glioma classification method newly released by WHO in 2021 to screen GBM patients into groups, and the brand new GBM classification and definition is more accurate for risk stratification of the patients. Secondly, the application performs multi-center verification based on the risk tag inside and outside, and the result is better consistency. Furthermore, the present application uses two transcriptome data analysis methods (GSEA and WGCNA) to explore the biological implications of the image histology features, which increases the robustness and convincing of the biological interpretation of the image histology features.
Referring to fig. 3, an embodiment of the present application also provides a system for risk stratification of GBM patients based on conventional MRI sequences, comprising: the system comprises a data set module 101, a conventional MRI sequence acquisition module 102, a model construction module 103 and a data analysis module 104 which are connected in sequence; the data set module 101 is configured to acquire a data set of a GBM patient; GBM patients are glioblastoma patients; the data set comprises a training set and a testing set; the conventional MRI sequence acquisition module 102 is configured to perform image acquisition on the data set to obtain a conventional MRI sequence; based on the data set, acquiring transcriptome sequencing data; conventional MRI sequences include fluid attenuation inversion recovery imaging sequences, T1 weighted contrast enhancement imaging sequences, T1 weighted imaging sequences, and T2 weighted imaging sequences; the model construction module 103 is used for extracting image histology characteristics according to a conventional MRI sequence; constructing an image clinical model based on the image histology characteristics; the data analysis module 104 is configured to obtain an image histology analysis result according to the image clinical model and the transcriptome sequencing data; analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; and explaining the biological basis behind the image group biological characteristics based on the image clinical model and the cross passage.
In some embodiments, referring still to fig. 3, the model building module 103 includes a preprocessing module 1031, a feature extraction module 1032, an image histology labeling module 1033, and an image clinical model module 1034 connected in sequence; the preprocessing module 1031 is configured to preprocess the conventional MRI sequence to obtain a preprocessed conventional MRI sequence; the feature extraction module 1032 is configured to extract image histology features from each three-dimensional volume of interest according to the preprocessed conventional MRI sequence; the three-dimensional volume of interest is a three-dimensional volume that generates a tumor through three-dimensional simulation; the image histology features include texture histology features, shape histology features, and intensity histology features; the image histology tag module 1033 is configured to screen out the best image histology features from the data set according to the image histology features to form an image histology tag; the image clinical model module 1034 is used for constructing an image clinical model according to the image histology tag.
Based on the above, the method and the system for risk stratification of GBM patients based on the conventional MRI sequence provided by the application firstly use a glioma classification method newly issued by WHO in 2021 to screen GBM patients into groups, and the brand-new GBM classification and definition are more accurate for risk stratification of patients. Secondly, the application performs multi-center verification based on the risk tag inside and outside, and the result is better consistency. Furthermore, the present application uses two transcriptome data analysis methods (GSEA and WGCNA) to explore the biological implications of the image histology features, which increases the robustness of the biological interpretation of the image histology features. Finally, we explore the biological meaning of the image histology features from two aspects. In one aspect, we elaborate on the number, type and distribution of biological pathways associated with a single image histology feature. On the other hand, we have also studied the biological interpretation of different DTI parameters.
By the above technical scheme, the embodiment of the application provides a method and a system for risk stratification of GBM patients based on a conventional MRI sequence, wherein the method comprises the following steps: firstly, acquiring a data set of a GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set; then, carrying out image acquisition on the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence; next, based on the conventional MRI sequence, extracting image histology features; constructing an image clinical model based on the image histology characteristics; then, based on the image clinical model and the transcriptome sequencing data, obtaining an image histology analysis result; next, analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; finally, the biological basis behind the image group biological characteristics is explained based on the image clinical model and the cross passage.
According to the method for risk stratification of the GBM patient based on the conventional MRI sequence, the risk stratification and prognosis prediction are carried out on the IDH wild Glioblastoma (GBM) patient by utilizing the image histology characteristics from the conventional MRI sequence, the biological meaning of the image histology characteristics is explored by adopting a reliable image genomics analysis technology, and the biological path behind the image histology prognosis characteristics is excavated, so that important guidance is provided for clinical accurate treatment.
That is, it will be understood by those skilled in the art that all or part of the steps in implementing the methods of the embodiments described above may be implemented by a program stored in a storage medium, where the program includes several instructions for causing a device (which may be a single-chip microcomputer, a chip or the like) or a processor (processor) to perform all or part of the steps in the methods of the embodiments described above. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific examples of carrying out the application and that various changes in form and details may be made therein without departing from the spirit and scope of the application. Various changes and modifications may be made by one skilled in the art without departing from the spirit and scope of the application, and the scope of the application is therefore intended to be limited only by the appended claims.
Claims (10)
1. A method of risk stratification for GBM patients based on conventional MRI sequences, comprising:
acquiring a data set of a GBM patient; the GBM patient is a glioblastoma patient; the data set comprises a training set and a testing set;
image acquisition is carried out on the data set to obtain a conventional MRI sequence; and based on the dataset, obtaining transcriptome sequencing data; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence;
extracting image histology features based on the conventional MRI sequence; constructing an image clinical model based on the image histology characteristics;
obtaining an image histology analysis result based on the image clinical model and the transcriptome sequencing data;
analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics;
and explaining the biological basis behind the image group biological characteristics based on the image clinical model and the cross passage.
2. The method of risk stratification of a GBM patient based on a conventional MRI sequence of claim 1, characterized by extracting an image histology feature based on the conventional MRI sequence; and constructing an image clinical model based on the image histology features, comprising:
Preprocessing the conventional MRI sequence to obtain a preprocessed conventional MRI sequence;
extracting image histology features from each three-dimensional volume of interest based on the preprocessed conventional MRI sequences; the three-dimensional interesting volume is a three-dimensional volume for generating tumors through three-dimensional simulation; the image histology features comprise texture histology features, shape histology features and intensity histology features;
screening out the optimal image histology characteristics from the data set based on the image histology characteristics to form an image histology tag;
and constructing an image clinical model based on the image histology tag.
3. The method of risk stratification for GBM patients based on conventional MRI sequences according to claim 2, characterized in that the imaging histology signature is used for predicting the prognosis of survival of IDH wild-type GBM patients.
4. The method of risk stratification for GBM patients based on conventional MRI sequences of claim 2 wherein screening the data set for the best image histology features based on the image histology features to constitute an image histology tag comprises:
calculating intra-group correlation coefficients of each image group learning feature, and screening out features with sketch consistency larger than a threshold value based on the intra-group correlation coefficients to obtain primary screening features;
Carrying out single factor regression analysis on the primary screening characteristics, and counting the consistency index of the primary screening characteristics to obtain secondary screening characteristics;
and screening the secondary screening characteristics again by adopting a minimum absolute shrinkage and selection operator screening method, and screening out the optimal image histology characteristics to form an image histology label.
5. The method of risk stratification for a GBM patient based on a conventional MRI sequence of claim 1, wherein constructing an imaging clinical model based on the imaging histology features comprises:
constructing an image group learning risk score based on the image group learning features;
incorporating clinical factors into a multi-factor regression analysis model, and establishing a clinical model for predicting total survival; the clinical factors include sex, age, surgical resection degree, whether chemoradiotherapy and preoperative Carlsberg score;
and adding the image group academic risk score into the clinical factors to obtain new clinical factors, and incorporating the new clinical factors into the multi-factor regression analysis model to construct an image clinical model.
6. The method of risk stratification for a GBM patient based on a conventional MRI sequence of claim 1, further comprising, after constructing an imaging clinical model based on the imaging histology characteristics, before deriving imaging histology analysis results based on the imaging clinical model and the transcriptome sequencing data:
And evaluating the image clinical model.
7. The method of risk stratification for a GBM patient based on a conventional MRI sequence of claim 6, wherein evaluating said imaging clinical model comprises:
and evaluating the image clinical model by adopting a survival analysis curve, a calibration curve and a decision curve respectively.
8. The method of risk stratification for a GBM patient based on a conventional MRI sequence of claim 1 wherein interpreting the biological basis behind an imaging histology feature based on the imaging clinical model and the crossover pathway comprises:
based on the image clinical model and the cross path, the biological basis under different risk stratification, the biological meaning of the single image group biological feature and the biological meaning of the single MRI sequence are respectively revealed to explain the biological basis behind the image group biological feature.
9. A system for risk stratification of GBM patients based on conventional MRI sequences, comprising: the system comprises a data set module, a conventional MRI sequence acquisition module, a model construction module and a data analysis module which are connected in sequence;
the data set module is used for acquiring a data set of the GBM patient; the GBM patient is a glioblastoma patient;
The data set comprises a training set and a testing set;
the conventional MRI sequence acquisition module is used for acquiring images of the data set to obtain a conventional MRI sequence; and is combined with
Based on the dataset, transcriptome sequencing data is acquired; the conventional MRI sequence includes a fluid attenuation inversion recovery imaging sequence, a T1 weighted contrast enhancement imaging sequence, a T1 weighted imaging sequence, and a T2 weighted imaging sequence;
the model construction module is used for extracting image histology characteristics according to the conventional MRI sequence; constructing an image clinical model based on the image histology characteristics;
the data analysis module is used for obtaining an image histology analysis result according to the image clinical model and the transcriptome sequencing data; analyzing transcriptome sequencing data by adopting a weighted gene co-expression network analysis and a gene set enrichment analysis method to obtain a cross path for revealing biological meanings of image histology characteristics; and explaining the biological basis behind the image group biological characteristics based on the image clinical model and the cross passage.
10. The system for risk stratification of a GBM patient based on a conventional MRI sequence of claim 9, wherein the model construction module comprises a preprocessing module, a feature extraction module, an image histology labeling module and an image clinical model module connected in sequence;
The pretreatment module is used for carrying out pretreatment on the conventional MRI sequence to obtain a pretreated conventional MRI sequence;
the feature extraction module is used for extracting image histology features from each three-dimensional interested volume according to the preprocessed conventional MRI sequences; the three-dimensional interesting volume is a three-dimensional volume for generating tumors through three-dimensional simulation; the image histology features comprise texture histology features, shape histology features and intensity histology features;
the image histology tag module is used for screening out the optimal image histology features from the data set according to the image histology features so as to form an image histology tag;
the image clinical model module is used for constructing an image clinical model according to the image histology label.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310667266.0A CN116705296A (en) | 2023-06-06 | 2023-06-06 | Method and system for risk stratification of GBM patient based on conventional MRI sequence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310667266.0A CN116705296A (en) | 2023-06-06 | 2023-06-06 | Method and system for risk stratification of GBM patient based on conventional MRI sequence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116705296A true CN116705296A (en) | 2023-09-05 |
Family
ID=87827126
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310667266.0A Pending CN116705296A (en) | 2023-06-06 | 2023-06-06 | Method and system for risk stratification of GBM patient based on conventional MRI sequence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116705296A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117116339A (en) * | 2023-10-24 | 2023-11-24 | 中日友好医院(中日友好临床医学研究所) | Method and device for identifying image group biological characteristics based on WGCNA |
| CN117173167A (en) * | 2023-11-02 | 2023-12-05 | 中日友好医院(中日友好临床医学研究所) | Method and device for predicting tumor prognosis by image histology machine learning survival model |
-
2023
- 2023-06-06 CN CN202310667266.0A patent/CN116705296A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117116339A (en) * | 2023-10-24 | 2023-11-24 | 中日友好医院(中日友好临床医学研究所) | Method and device for identifying image group biological characteristics based on WGCNA |
| CN117173167A (en) * | 2023-11-02 | 2023-12-05 | 中日友好医院(中日友好临床医学研究所) | Method and device for predicting tumor prognosis by image histology machine learning survival model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2739675C (en) | Gene and gene expressed protein targets depicting biomarker patterns and signature sets by tumor type | |
| CN116705296A (en) | Method and system for risk stratification of GBM patient based on conventional MRI sequence | |
| CN1385702A (en) | Method for supply clinical diagnosis | |
| CN107102151A (en) | Cardiovascular danger event prediction and application thereof | |
| JP2021505977A (en) | Methods and systems for determining somatic mutation clonality | |
| Kumar et al. | Biomarkers of diseases in medicine | |
| CN107408163A (en) | For analyzing the method and device of gene | |
| US20190018930A1 (en) | Method for building a database | |
| Reggiardo et al. | LncRNA biomarkers of inflammation and cancer | |
| Connell et al. | Development of a multivariable risk model integrating urinary cell DNA methylation and cell‐free RNA data for the detection of significant prostate cancer | |
| WO2021081253A1 (en) | Systems and methods for predicting therapeutic sensitivity | |
| Webster et al. | Quantifying histological features of cancer biospecimens for biobanking quality assurance using automated morphometric pattern recognition image analysis algorithms | |
| Katrib et al. | “RADIOTRANSCRIPTOMICS”: a synergy of imaging and transcriptomics in clinical assessment | |
| Meng et al. | The application of machine learning techniques in clinical drug therapy | |
| Burstein et al. | Genetic variant burden and adverse outcomes in pediatric cardiomyopathy | |
| CN116385441B (en) | Method and system for risk stratification of oligodendroglioma based on MRI | |
| Wedderburn et al. | Should all individuals be screened for genetic predisposition to cancer? | |
| Deshpande et al. | Relationship between donor fraction cell‐free DNA and clinical rejection in heart transplantation | |
| CN107408162A (en) | For analyzing the method and device of gene | |
| CN116403076B (en) | Method and system for risk stratification of GBM patient based on DTI sequence | |
| Hobbs et al. | Biostatistics and bioinformatics in clinical trials | |
| KR102161511B1 (en) | Extracting method for biomarker for diagnosis of biliary tract cancer, computing device therefor, biomarker for diagnosis of biliary tract cancer, and biliary tract cancer diagnosis device comprising same | |
| KR102138517B1 (en) | Extracting method for biomarker for diagnosis of pancreatic cancer, computing device therefor, biomarker, and pancreatic cancer diagnosis device comprising same | |
| Yuan et al. | Convolutional neural network to predict IDH mutation status in glioma from chemical exchange saturation transfer imaging at 7 Tesla | |
| US20230289569A1 (en) | Non-Transitory Computer Readable Medium, Information Processing Device, Information Processing Method, and Method for Generating Learning Model |
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 |