CN117174323A - SFTs integration risk assessment system - Google Patents
SFTs integration risk assessment system Download PDFInfo
- Publication number
- CN117174323A CN117174323A CN202311186054.7A CN202311186054A CN117174323A CN 117174323 A CN117174323 A CN 117174323A CN 202311186054 A CN202311186054 A CN 202311186054A CN 117174323 A CN117174323 A CN 117174323A
- Authority
- CN
- China
- Prior art keywords
- sfts
- risk assessment
- features
- model
- variation
- 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
- 238000012502 risk assessment Methods 0.000 title claims abstract description 141
- 230000010354 integration Effects 0.000 title claims description 4
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 54
- 230000008595 infiltration Effects 0.000 claims abstract description 42
- 238000001764 infiltration Methods 0.000 claims abstract description 42
- 230000004083 survival effect Effects 0.000 claims abstract description 39
- 230000001575 pathological effect Effects 0.000 claims abstract description 38
- 230000007614 genetic variation Effects 0.000 claims abstract description 36
- 108090000623 proteins and genes Proteins 0.000 claims abstract description 9
- 101000772904 Homo sapiens Ubiquitin-conjugating enzyme E2 D1 Proteins 0.000 claims description 117
- 102100030433 Ubiquitin-conjugating enzyme E2 D1 Human genes 0.000 claims description 104
- 238000012549 training Methods 0.000 claims description 35
- 230000007170 pathology Effects 0.000 claims description 29
- 108010074708 B7-H1 Antigen Proteins 0.000 claims description 17
- 108010009992 CD163 antigen Proteins 0.000 claims description 17
- 101150097381 Mtor gene Proteins 0.000 claims description 17
- 102100025831 Scavenger receptor cysteine-rich type 1 protein M130 Human genes 0.000 claims description 17
- 230000017074 necrotic cell death Effects 0.000 claims description 10
- 102100023085 Serine/threonine-protein kinase mTOR Human genes 0.000 claims description 9
- 102100025064 Cellular tumor antigen p53 Human genes 0.000 claims description 8
- 108010078814 Tumor Suppressor Protein p53 Proteins 0.000 claims description 8
- 230000011278 mitosis Effects 0.000 claims description 8
- 108700039691 Genetic Promoter Regions Proteins 0.000 claims description 7
- 102100036011 T-cell surface glycoprotein CD4 Human genes 0.000 claims description 7
- 102100022005 B-lymphocyte antigen CD20 Human genes 0.000 claims description 6
- 101001042041 Bos taurus Isocitrate dehydrogenase [NAD] subunit beta, mitochondrial Proteins 0.000 claims description 6
- 102100031866 DNA excision repair protein ERCC-5 Human genes 0.000 claims description 6
- 108010035476 DNA excision repair protein ERCC-5 Proteins 0.000 claims description 6
- 102100031618 HLA class II histocompatibility antigen, DP beta 1 chain Human genes 0.000 claims description 6
- 108010045483 HLA-DPB1 antigen Proteins 0.000 claims description 6
- 101000897405 Homo sapiens B-lymphocyte antigen CD20 Proteins 0.000 claims description 6
- 101000960234 Homo sapiens Isocitrate dehydrogenase [NADP] cytoplasmic Proteins 0.000 claims description 6
- 101000934372 Homo sapiens Macrosialin Proteins 0.000 claims description 6
- 102100022297 Integrin alpha-X Human genes 0.000 claims description 6
- 102100039905 Isocitrate dehydrogenase [NADP] cytoplasmic Human genes 0.000 claims description 6
- 102100025136 Macrosialin Human genes 0.000 claims description 6
- 102100034922 T-cell surface glycoprotein CD8 alpha chain Human genes 0.000 claims description 6
- 102000017420 CD3 protein, epsilon/gamma/delta subunit Human genes 0.000 claims description 5
- 108050005493 CD3 protein, epsilon/gamma/delta subunit Proteins 0.000 claims description 5
- 102100027581 Forkhead box protein P3 Human genes 0.000 claims description 5
- 101000861452 Homo sapiens Forkhead box protein P3 Proteins 0.000 claims description 5
- 101000716102 Homo sapiens T-cell surface glycoprotein CD4 Proteins 0.000 claims description 5
- 101000946843 Homo sapiens T-cell surface glycoprotein CD8 alpha chain Proteins 0.000 claims description 5
- 102000001759 Notch1 Receptor Human genes 0.000 claims description 5
- 108010029755 Notch1 Receptor Proteins 0.000 claims description 5
- 238000002512 chemotherapy Methods 0.000 claims description 5
- 230000019491 signal transduction Effects 0.000 claims description 5
- 206010033661 Pancytopenia Diseases 0.000 claims description 4
- 208000024389 cytopenia Diseases 0.000 claims description 4
- -1 ki-67 Proteins 0.000 claims description 4
- 230000005855 radiation Effects 0.000 claims description 3
- 102100024216 Programmed cell death 1 ligand 1 Human genes 0.000 claims 1
- 238000004393 prognosis Methods 0.000 abstract description 24
- 238000000034 method Methods 0.000 abstract description 10
- 230000003176 fibrotic effect Effects 0.000 abstract description 2
- 230000000750 progressive effect Effects 0.000 abstract 2
- 208000014653 solitary fibrous tumor Diseases 0.000 description 110
- 210000004027 cell Anatomy 0.000 description 41
- 210000003169 central nervous system Anatomy 0.000 description 17
- 102000008096 B7-H1 Antigen Human genes 0.000 description 16
- 210000001519 tissue Anatomy 0.000 description 16
- 230000014509 gene expression Effects 0.000 description 13
- 238000012795 verification Methods 0.000 description 11
- 206010027476 Metastases Diseases 0.000 description 9
- 206010028851 Necrosis Diseases 0.000 description 9
- 210000002865 immune cell Anatomy 0.000 description 9
- 230000009401 metastasis Effects 0.000 description 8
- 238000003745 diagnosis Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 210000004881 tumor cell Anatomy 0.000 description 6
- 210000003484 anatomy Anatomy 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 5
- 230000003211 malignant effect Effects 0.000 description 5
- 229920002153 Hydroxypropyl cellulose Polymers 0.000 description 4
- 201000007455 central nervous system cancer Diseases 0.000 description 4
- 208000025997 central nervous system neoplasm Diseases 0.000 description 4
- 235000010977 hydroxypropyl cellulose Nutrition 0.000 description 4
- 238000012216 screening Methods 0.000 description 4
- 238000013517 stratification Methods 0.000 description 4
- 238000001356 surgical procedure Methods 0.000 description 4
- 102000008186 Collagen Human genes 0.000 description 3
- 108010035532 Collagen Proteins 0.000 description 3
- 208000006011 Stroke Diseases 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 3
- 229920001436 collagen Polymers 0.000 description 3
- 230000034994 death Effects 0.000 description 3
- 238000010348 incorporation Methods 0.000 description 3
- 239000003550 marker Substances 0.000 description 3
- 238000001959 radiotherapy Methods 0.000 description 3
- 238000012163 sequencing technique Methods 0.000 description 3
- 102100023181 Neurogenic locus notch homolog protein 1 Human genes 0.000 description 2
- 108010011005 STAT6 Transcription Factor Proteins 0.000 description 2
- 102000013968 STAT6 Transcription Factor Human genes 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 210000002540 macrophage Anatomy 0.000 description 2
- 210000004379 membrane Anatomy 0.000 description 2
- 239000012528 membrane Substances 0.000 description 2
- 230000000394 mitotic effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 208000015356 pericytic neoplasm Diseases 0.000 description 2
- 201000008944 perivascular tumor Diseases 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 230000035755 proliferation Effects 0.000 description 2
- 238000004445 quantitative analysis Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000011282 treatment Methods 0.000 description 2
- 206010006187 Breast cancer Diseases 0.000 description 1
- 208000026310 Breast neoplasm Diseases 0.000 description 1
- 102100021975 CREB-binding protein Human genes 0.000 description 1
- 206010009944 Colon cancer Diseases 0.000 description 1
- 208000001333 Colorectal Neoplasms Diseases 0.000 description 1
- 206010011224 Cough Diseases 0.000 description 1
- 206010061818 Disease progression Diseases 0.000 description 1
- 208000000059 Dyspnea Diseases 0.000 description 1
- 206010013975 Dyspnoeas Diseases 0.000 description 1
- 101000896987 Homo sapiens CREB-binding protein Proteins 0.000 description 1
- 206010020877 Hypertrophic osteoarthropathy Diseases 0.000 description 1
- 230000005913 Notch signaling pathway Effects 0.000 description 1
- 102000001756 Notch2 Receptor Human genes 0.000 description 1
- 102000001760 Notch3 Receptor Human genes 0.000 description 1
- 208000028895 Potter sequence Diseases 0.000 description 1
- 208000019552 Potter syndrome Diseases 0.000 description 1
- 208000035582 Rare soft tissue tumor Diseases 0.000 description 1
- 210000001744 T-lymphocyte Anatomy 0.000 description 1
- 206010054094 Tumour necrosis Diseases 0.000 description 1
- 210000000683 abdominal cavity Anatomy 0.000 description 1
- 210000003719 b-lymphocyte Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000004791 biological behavior Effects 0.000 description 1
- 230000000740 bleeding effect Effects 0.000 description 1
- 238000007475 c-index Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 239000002771 cell marker Substances 0.000 description 1
- 210000003855 cell nucleus Anatomy 0.000 description 1
- 210000000038 chest Anatomy 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 210000004443 dendritic cell Anatomy 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 230000005750 disease progression Effects 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 210000002950 fibroblast Anatomy 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000002218 hypoglycaemic effect Effects 0.000 description 1
- 229940088592 immunologic factor Drugs 0.000 description 1
- 239000000367 immunologic factor Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007912 intraperitoneal administration Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 210000004072 lung Anatomy 0.000 description 1
- 210000004698 lymphocyte Anatomy 0.000 description 1
- 230000036210 malignancy Effects 0.000 description 1
- 210000001370 mediastinum Anatomy 0.000 description 1
- 210000002418 meninge Anatomy 0.000 description 1
- 201000008806 mesenchymal cell neoplasm Diseases 0.000 description 1
- 206010061289 metastatic neoplasm Diseases 0.000 description 1
- 210000004088 microvessel Anatomy 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 210000000651 myofibroblast Anatomy 0.000 description 1
- 210000004940 nucleus Anatomy 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 239000012188 paraffin wax Substances 0.000 description 1
- 208000012111 paraneoplastic syndrome Diseases 0.000 description 1
- 238000010827 pathological analysis Methods 0.000 description 1
- 210000003668 pericyte Anatomy 0.000 description 1
- 210000004224 pleura Anatomy 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000002685 pulmonary effect Effects 0.000 description 1
- 238000011127 radiochemotherapy Methods 0.000 description 1
- 235000008001 rakum palm Nutrition 0.000 description 1
- 238000007480 sanger sequencing Methods 0.000 description 1
- 210000002460 smooth muscle Anatomy 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 208000011580 syndromic disease Diseases 0.000 description 1
- 238000002626 targeted therapy Methods 0.000 description 1
- 238000011269 treatment regimen Methods 0.000 description 1
- 230000009452 underexpressoin Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 210000003556 vascular endothelial cell Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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 invention discloses an integrated risk assessment system for isolated fibrotic tumors (SFTs), which comprises the following components: the system comprises a feature selection module, a risk assessment module and a progression-free survival assessment module; acquiring selected pathological features, selected immune infiltration features and selected gene variation information corresponding to a patient to be detected through a feature selection module, inputting the selected pathological features, the selected immune infiltration features and the selected gene variation information into a risk assessment model through a risk assessment module to determine risk assessment scores of the corresponding patient, and assessing survival probability of the patient to be detected in each non-progressive lifetime according to the risk assessment scores through a non-progressive survival assessment module; by implementing the method, the prognosis of the SFTs patient to be detected can be predicted from three dimensions of pathological features, immune infiltration features and genetic variation features, and compared with the existing 2020WHO classification, mDemicco model and G-score model, the accuracy of predicting the prognosis of the SFTs patient to be detected is improved.
Description
Technical Field
The invention relates to the field of SFTs prognosis prediction, in particular to an SFTs integration risk assessment system.
Background
Isolated fibrotic tumors (Solitary fibrous tumors, SFTs) are a rare mesenchymal tumor that originates from fibroblasts/myofibroblasts. SFTs are common in adults, with peaks of onset between 50-60 years, and can occur in any anatomical part of the body, with the chest (pleura, mediastinum, lung parenchyma) being the most common first place, and the most common extrapleural parts being the head (including meninges) and neck, and then the abdominal cavity (peritoneal, retroperitoneal), trunk and extremities. SFTs are often slow growing, patients may develop non-specific pulmonary symptoms such as dyspnea or cough, tumor volumes are large enough to press against adjacent anatomy, and in addition, few patients may develop paraneoplastic syndrome, hypertrophic osteoarthropathy, and less than 5% of SFTs patients may also develop refractory hypoglycemic syndrome, known as Doege-Potter syndrome.
One of the main problems occurring in clinical diagnosis and treatment of SFTs is: the existing benign and malignant hierarchical diagnosis standard cannot accurately predict malignant potential. The histology of SFTs is broad, morphologically representing few-to multicellular tumors, consisting of atypical spindle cells, with nuclei irregularly arranged in spindle form, surrounded by dense interstitial collagen and thin collagen bands, and mixed with typical deer-horn branched vessels (vascular endothelial cell tumor-like), but these histopathological patterns are not characteristic of SFTs, but can be observed in other mesenchymal tumors. Meningeal SFTs, previously known as perivascular tumors (HPCs), originate from the endocranial microvessels surrounding smooth muscle pericytes, and have been considered as low-grade tumors, although the clinical manifestations are significantly different, in the fourth WHO classification, SFTs and HPCs were combined in one disease classification, the fourth revision of WHO central nervous system (Central Nervous System, CNS) tumor classification published in 2016 introduced the mixed "SFTs/HPCs" classification and were subdivided into three histological grades according to the division count, the term "perivascular tumor" was deleted in the 2021WHO central nervous system tumor classification. The fifth version of WHO classification published in month 4 2020 regards SFTs and HPCs as the same entity at both ends of the same tissue spectrum and classifies them into three categories: benign (locally invasive), unspecified and malignant. Although WHO classification is continually updated as the awareness of SFTs increases, the criteria for classifying morphologically benign and malignant SFTs is not able to accurately predict the malignant potential and biological behavior of SFTs, patients with morphologically benign manifestations can still progress, and about 10% -40% of SFTs relapse or metastasis occurs, and most often in the first 5 years after diagnosis, but there are also literature reports that tumor recurrence can still occur in the last decade. Therefore, a new SFTs risk assessment model is urgently needed to be provided for accurately risk stratification of SFTs, so as to guide clinic to accurately formulate a treatment strategy of the SFTs.
In order to accurately stratify risk for patients with SFTs, a number of risk assessment models composed of a variety of prognostic factors (mainly clinical and histopathological factors) have been proposed in recent years based on the original morphological diagnosis to predict risk of relapse/metastasis in individuals, wherein the most accepted models are mdemico model and G-score model. In addition, there are also established, with high acceptance, the Pasquali model (nuclear division count, cytorich, nuclear polymorphism is incorporated into the established model) based on pleural SFTs, the Salas OS proposed based on non-central nervous system SFTs (incorporation variables include: age, nuclear division count), salas MET (incorporation variables include: age, nuclear division count, tumor site), and the Salas LR model (incorporation variables include: age, tumor site, whether or not radiotherapy), and the like.
The Demicco model is an SFTs risk assessment model proposed by the professor Elizabeth G Demicco of the medical science of Kaishan Hospital, new York, U.S. and revised in 2017, simply called mDemicco model, which incorporates patient age (< 55 years 0 min, > 55 years 1 min), tumor size (< 5cm 0 min, 5-10cm1 min, 10-15cm 2 min, > 15cm 3 min), mitosis count (0/10 HPF 0 min, 1-3/10HPF 1 min, > 4/10HPF 2 min) and necrosis (< 10%0 min, > 10%1 min), four clinical pathology parameters, four cumulative scores of 0-3 min into low risk groups, 4-5 min into stroke risk groups, and 6-7 min into high risk groups. The range of application of the mdemico model is: primary non-central nervous system SFTs. The prognostic indicator of this model is Metastasis-free survival (MFS), defined as the time from surgery to tumor Metastasis (excluding primary foci recurrence and death).
Professor Tatiana Georgiesh to Norway pathology in 2020 published in Histoparology on the G-score model. G-score eventually incorporates mitotic counts (< 4/10HPF 0 score, > 4/10HPF 2 score), necrosis (no score of 0, < 50%1 score, > 50%2 score), gender (female 0 score, male 1 score), three cumulative scores of 0 divided into low risk groups, 1-2 divided into stroke risk groups, and 3-5 divided into high risk groups. The inclusion criteria for the G-score model were: non-central nervous system, STAT6 positive primary lesions, except specimens that received preoperatively chemoradiotherapy. The prognostic indicator is a recurrence-free interval (RFi), defined as the time from surgery to tumor recurrence (including distant metastasis and local recurrence, excluding death). The greatest advantage of this model is the long follow-up time, median OS (Overall survival) and RFi of 121 and 84 months respectively, which can be used to predict short-term and long-term recurrence.
Although these risk assessment models can predict clinical progression of SFTs to some extent, there is a continuing literature indicating that they cannot accurately predict malignancy potential of SFTs, and there is still significant room for improvement. Furthermore, the mdemico model is not applicable to SFTs of the central nervous system and is based on western population studies. The inclusion of genetic variation information in tumors such as colorectal cancer, breast cancer and the like can significantly improve the performance of a clinical prognosis model. NAB2-STAT6 fusion genes and STAT6 immune expression widely studied in SFTs tissues are only used as diagnostic indexes and are irrelevant to the prognosis of the SFTs. Recent studies report that incorporating TERT promoter region variation, TP53 variation, into a prognostic model may increase the predictive accuracy of the model. Unfortunately, there is no study currently that integrates genetic variation information of SFTs into a risk assessment model, nor that incorporates SFTs immunoinfiltration analysis into a risk assessment model. That is, the commonly used risk assessment model is used for predicting prognosis of the SFTs patient only through the single dimension of clinical pathological parameters, and the prognosis prediction result of the SFTs patient is not accurate enough.
Disclosure of Invention
The invention provides an SFTs integrated risk assessment system, which can improve the accuracy of SFTs prognosis prediction compared with 2020WHO classification, mDemicco model and G-score model.
The invention provides an SFTs integrated risk assessment system, which comprises: the system comprises a feature selection module, a risk assessment module and a progression-free survival assessment module;
the characteristic selection module is used for acquiring selected pathological characteristics, selected immune infiltration characteristics and selected genetic variation information corresponding to a patient to be detected;
the risk assessment module is used for inputting the pathological features, the immune infiltration features and the genetic variation information into a risk assessment model so that the risk assessment model determines a risk assessment score according to the pathological features, the immune infiltration features and the genetic variation information;
the non-progress survival assessment module is used for assessing the survival probability of the patient to be detected in each non-progress survival period according to the risk assessment score;
the risk assessment model is obtained by taking selected pathological features, selected immune infiltration features and selected genetic variation information of training samples as input and taking risk assessment scores corresponding to the training samples as output.
Further, the selected pathological feature comprises: a mitosis count; the selected immunoinfiltration features include: CD163 positive cell density and Ki-67 positive cell density; the selected genetic variation information comprises: MTOR gene variation information.
Further, the method further comprises the following steps:
and constructing an alignment chart corresponding to the risk assessment model according to the selected pathological features, the selected immune infiltration features, the selected genetic variation information, the risk assessment score and the probability of each progression-free survival period.
Further, in training the risk assessment model, selected pathological features, selected immune infiltration features, and selected genetic variation information of the training sample are determined by:
acquiring an SFTs clinical pathological feature set, an SFTs gene variation feature set and an SFTs immune infiltration feature set;
according to a preset LASSO regression model, a first training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
according to a preset random survival forest model, a second training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
and constructing an intersection of the first training subset and the second training subset, and determining selected pathological features, selected immune infiltration features and selected genetic variation information of the training sample according to the intersection.
Further, the set of SFTs clinical pathology features includes: tumor site, patient age, patient sex, tumor size, nuclear division count, necrosis, nuclear polymorphism, tumor tissue cytopenia, patient radiation information, and patient chemotherapy information.
Further, the SFTs gene variation feature set includes: MTOR variation, NOTCH1 signaling pathway variation, ERCC5 variation, TP53 variation, TERT promoter region variation, and IDH1 variation.
Further, the SFTs immunoinfiltration feature set includes: densities of PD-L1, ki-67, CD68, CD163, HLA-DPB1, CD3, CD4, CD8, FOXP3, CD11c and CD20 positive cells.
Further, the method further comprises the following steps: a risk division module;
the risk classification module is used for classifying the patients to be detected into a low risk group, a medium risk group or a high risk group according to the risk assessment score;
wherein when the risk assessment score is less than or equal to a first threshold, dividing the patient to be detected into a low risk group; dividing the patient to be detected into risk groups when the risk assessment score is greater than a first threshold and less than or equal to a second threshold; dividing the patient to be detected into high-risk groups when the risk assessment score is greater than a second threshold value; the second threshold is greater than the first threshold.
The embodiment of the invention has the following beneficial effects:
the invention provides an SFTs integrated risk assessment system, which comprises: the system comprises a feature selection module, a risk assessment module and a progression-free survival assessment module; according to the invention, the selected pathological features, the selected immune infiltration features and the selected genetic variation information corresponding to the patient to be detected are input into the risk assessment model, and the risk assessment score is determined through the risk assessment model, so that the survival probability of the patient to be detected in each progression-free lifetime is determined according to the risk assessment score. In addition to analyzing the influence of the SFTs clinical pathology characteristics on prognosis, the invention analyzes the influence of the SFTs immune infiltration characteristics and genetic variation on SFTs prognosis, explores factors possibly influencing the SFTs prognosis from a multidimensional aspect and brings the factors into a risk assessment model, and compared with 2020WHO classification, mDemicco model and G-score model, the invention improves the accuracy of predicting the SFTs prognosis and solves the problem that the existing model only brings the clinical pathology characteristics into the SFTs prognosis and can not accurately predict the SFTs prognosis.
Drawings
FIG. 1 is a schematic diagram of an SFTs integrated risk assessment system according to an embodiment of the present invention;
FIG. 2 is a schematic view of a risk assessment model according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a feature screening method according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating verification of a risk assessment model in a verification queue according to an embodiment of the present invention;
FIG. 5 is a schematic diagram showing a comparison of a risk assessment model and an existing model according to an embodiment of the present invention;
FIG. 6 is a schematic diagram showing a comparison of another risk assessment model with an existing model according to an embodiment of the present invention;
FIG. 7 is a schematic representation of a risk assessment model in the primary CNS according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made more apparent and fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments of the invention are shown. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, an SFTs integrated risk assessment system according to an embodiment includes: the system comprises a feature selection module, a risk assessment module and a progression-free survival assessment module;
the characteristic selection module is used for acquiring selected pathological characteristics, selected immune infiltration characteristics and selected gene variation characteristics corresponding to a patient to be detected;
the risk assessment module is used for inputting the pathological features, the immune infiltration features and the genetic variation information into a risk assessment model so that the risk assessment model determines a risk assessment score according to the pathological features, the immune infiltration features and the genetic variation features;
the non-progress survival assessment module is used for assessing the survival probability of the patient to be detected in each non-progress survival period according to the risk assessment score;
the risk assessment model is obtained by taking selected pathological features, selected immune infiltration features and selected genetic variation features of training samples as inputs and taking risk assessment scores corresponding to the training samples as outputs.
Specifically, firstly, through a feature selection module, selected pathological features, selected immune infiltration features and selected genetic variation information corresponding to a user to be detected are obtained; then inputting the acquired pathological features, immune infiltration features and genetic variation features into a risk assessment model through a risk assessment module, enabling the risk assessment model to determine risk assessment scores according to the input pathological features, immune infiltration features and genetic variation information, and finally assessing the survival probability of a patient to be detected in each progression-free survival period according to the determined risk assessment scores through a progression-free survival assessment module; the risk assessment model is obtained by training selected pathological features, selected immune infiltration features and selected genetic variation features of training samples as inputs and the risk assessment score corresponding to each training sample as output, and compared with the existing 2020WHO classification, mdemico model and G-score model, the integrated risk assessment model is better and verified in three independent verification queues.
In a preferred embodiment, the set of SFTs clinical pathology features includes, but is not limited to: tumor site, patient age, patient sex, tumor size, nuclear division count, necrosis, nuclear polymorphism, tumor tissue cytopenia, patient radiation information, and patient chemotherapy information.
The method comprises the steps of collecting 101 cases of paraffin embedded tumor tissue specimens of SFTs of university tumor prevention centers in the middle mountain in 2008 to 2020, wherein the case inclusion standard is as follows, a, all tumors are proved to be SFTs through histopathology; b. all tumor specimens did not receive any chemotherapy, radiotherapy or targeted therapy prior to surgery; c. all patients had complete clinical pathology data; d, excluding recurrent metastasis SFTs; e. except for external border positive SFTs, the study was approved by the center ethical committee for tumor control at Zhongshan university (SYSUCC: B2021-421-01).
Acquiring clinical pathology information of a patient from a central electronic medical record system of a tumor prevention center of Zhongshan university; acquiring patient gender, age at diagnosis, specimen type, tumor size, tumor anatomical part and follow-up information by retrieving patient medical records; specimen type: primary tumor, recurrent or metastatic tumor; age at patient diagnosis: not less than 55 years old or less than 55 years old; tumor size, measuring the maximum diameter of tumor tissue of a surgical excision specimen, wherein the maximum diameter is <5cm or 5-10cm or 10-15cm or more than or equal to 15cm; tumor anatomical sites are divided into intrathoracic, head and neck, trunk, extremities, intraperitoneal and central nervous system; the follow-up mode is mainly electronic medical records of a medical record system, an imaging examination report and telephone follow-up, and the main end point of evaluation is Progression-free survival (PFS), which is defined as the time from surgery to disease Progression (recurrence or metastasis) or death. And analyzing the relation between the clinical pathological characteristics and the prognosis of the patient.
According to fifth edition of WHO classification released in month 4 2020, all HE sections of individual specimens were subjected to pathological diagnosis and reclassification. The following histopathological data were collected for each SFTs tissue: a nuclear division count of <4/10HPF or ≡4/10HPF per 10 High Power Field (HPF, x 400) of the most nuclear division active area of the tumor cells; necrosis, with the exception of bleeding or transparentized areas, tumor necrosis involving 10% of the tumor area or more is defined as the presence of necrosis; nuclear polymorphism, classified as low (cell monomorphic, with uniform nuclear characteristics), moderate (increased nuclear polymorphism, more prominent nucleoli and rare multi-nuclei cells) or high (nuclear deep staining, foci with apparent polymorphism and singular cells); tumor tissue is cell-rich, and is divided into low (tumor is composed mainly of hardened collagen with scattered, compressed spindle cells), moderate (many areas of increased cell density, cells adjacent to each other) and high (cell-rich tumor, cell-nucleus overlap). And analyzing the relation between the clinical pathological characteristics and prognosis of the patient.
In a preferred embodiment, the SFTs gene variation feature set includes, but is not limited to: MTOR variation, NOTCH1 signaling pathway variation, ERCC5 variation, TP53 variation, TERT promoter region variation, and IDH1 variation.
And detecting 1021 tumor-related gene variations of 101 SFT tissues by using a high-throughput targeted gene sequencing technology platform, and constructing a tumor gene variation map of the SFT tissues. The effect of MTOR variation, ERCC5 variation, TP53 variation, TERT promoter region variation and IDH1 variation or NOTCH1 signaling pathway gene variation on SFTs prognosis was analyzed. Particularly, there are many reported TP53 in the literature.
In a preferred embodiment, the SFTs immunoinfiltration feature set includes, but is not limited to: densities of PD-L1, ki-67, CD68, CD163, HLA-DPB1, CD3, CD4, CD8, FOXP3, CD11c and CD20 positive cells.
Detecting the expression of macrophage markers (CD 68, CD163, HLA-DPB 1), T cell markers (CD 3, CD4, CD8, FOXP 3), dendritic cell markers (CD 11 c), immune cell markers such as B cell marker CD20, immune checkpoint molecule PD-L1 and proliferation marker Ki-67 in SYSUCC cohort clinical samples using IHC; all sections were scanned by a Zeiss digital pathology slide scanner to generate digital pathology sections, and the HALO 2.3 digital pathology system quantitatively analyzed for CD68, CD163, HLA-DPB1, CD3, CD4, CD8, FOXP3, CD11c, and Ki-67 positive cell density (cells/mm) 2 ) The method comprises the steps of carrying out a first treatment on the surface of the The optimal cut-off value is defined by X-tile software, and all cases with the average density of positive cells being greater than the cut-off value are classified as high expression of the specific marker, and all cases with the average density of positive cells being lower than the cut-off value are classified as low expression of the specific marker; the distribution density of each immune cell was analyzed for relationship to the prognosis of patient survival.
HALO digital pathology quantitative analysis combined with visual observation under a microscope assessed PD-L1 expression in 101 SFTs tissues. The PD-L1 positive cell density was first analyzed by the hao 2.3 digital pathology system, and then the cell type of the PD-L1 positive cells was microscopically differentiated by naked eye as tumor cells or immune cells. For cases where tumor cells express PD-L1, the ratio of PD-L1 positive cells (total number of PD-L1 membrane stained cells/total number of all cells, including tumor cells, macrophages or lymphocytes, 100%) is ∈ 20% or more defined as SFTs with high expression of tumor cells PD-L1; for cases where the immune cells express PD-L1, the proportion of positive immune cells (total number of PD-L1 membrane stained immune cells/total immune cell count 100%) was defined as SFTs with high expression of PD-L1 immune cells; and analyzing the relation between the PD-L1 highly expressed by the SFTs tumor cells or immune cells and the survival prognosis and clinical pathological characteristics of the patients.
In a preferred embodiment, the selected pathological features include, but are not limited to: a mitosis count; the selected immunoinfiltration features include, but are not limited to: CD163 positive cell density and Ki-67 positive cell density; the selected genetic variation information includes, but is not limited to: MTOR gene variation information.
In a preferred embodiment, the selected pathology features, selected immunoinfiltration features, and selected genetic variation information of the training sample are determined in training the risk assessment model by:
acquiring an SFTs clinical pathological feature set, an SFTs gene variation feature set and an SFTs immune infiltration feature set;
according to a preset LASSO regression model, a first training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
according to a preset random survival forest model, a second training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
and constructing an intersection of the first training subset and the second training subset, and determining selected pathological features, selected immune infiltration features and selected genetic variation information of the training sample according to the intersection.
Specifically, after acquiring an SFTs clinical pathology feature set, an SFTs gene variation feature set and an SFTs immune infiltration feature set, screening and sorting are firstly performed by using a LASSO regression model, and a first training subset screened by the LASSO regression model is acquired. And screening and sequencing the importance of the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set through a random survival forest model to obtain a second training subset. And determining selected pathological features, selected immune infiltration features and selected genetic variation features of the training sample according to the intersection of the acquired first training subset and the second training subset.
In an alternative embodiment, SFTs clinical pathology (tumor site, age, sex, tumor size, nuclear division count, necrosis, nuclear polymorphism, tumor tissue richness, radiotherapy, chemotherapy), genetic variation information (including MTOR variation, NOTCH1 signaling pathway variation (NOTCH 1, NOTCH2, NOTCH3, ERBBP), ERCC5 variation, TP53 variation, TERT promoter region variation, and IDH1 variation) and immunoinfiltration characteristics (including CD68, CD163, HLA-DPB1, CD3, CD4, CD8, FOXP3, CD11c, CD20 positive cell density), ki-67 positive cell density, and PD-L1 expression level) are screened using LASSO regression, random survival forest, and the variables predicted by the two methods are sequentially added to the model in order of importance until the addition of variables fails to improve model performance, the evaluation criteria being a base score, which is one of the prior art; four variables of final nuclear division count, CD163 positive cell density, ki-67 positive cell density, and MTOR gene variation were included to construct an integrated risk assessment model.
In a preferred embodiment, the nomogram corresponding to the risk assessment model is constructed according to the selected pathology features, the selected immune infiltration features, the selected genetic variation features, the risk assessment score and the probability of each progression free survival.
Specifically, four variables selected by the feature selection module, MTOR gene variation, CD163 positive cell density, ki-67 positive cell density, and mitosis count, were included in the construction model, and the model was constructed and visualized using Nomogram, wherein the mitosis count was determined by<4/10HPF score is 0 score, and more than or equal to 4/10HPF score is 53 score; the wild type of the MTOR gene is 0 score, and the MTOR gene variation is 100 score; ki-67 positive cell density was lower than 454.7cells/mm 2 Score 0 score, higher than 454.7cells/mm 2 A score of 21; CD163 positive cell density was lower than 929.3cells/mm 2 Score 0 score, higher than 929.3cells/mm 2 Score 45; the total score of the four scores is the Nomogram total score of each SFTs, and the Nomogram total score is the risk assessment score;
in a preferred embodiment, the SFTs integrated risk assessment system further comprises: a risk division module; the risk classification module is used for classifying the patients to be detected into a low risk group, a medium risk group or a high risk group according to the risk assessment score;
wherein when the risk assessment score is less than or equal to a first threshold, dividing the patient to be detected into a low risk group; dividing the patient to be detected into risk groups when the risk assessment score is greater than a first threshold and less than or equal to a second threshold; dividing the patient to be detected into high-risk groups when the risk assessment score is greater than a second threshold value; the second threshold is greater than the first threshold.
Specifically, the first threshold includes, but is not limited to: 0; the second threshold includes, but is not limited to: 74; dividing the patients into a high risk group, a medium risk group and a low risk group according to the risk assessment score, wherein the risk assessment score is less than or equal to 0 for the low risk group; a score of 0< risk assessment score ∈74 for the medium risk group and a total score of Nomogram >74 for the high risk group.
In an alternative embodiment, integrating the evaluation, external verification and comparison of the risk assessment model includes:
a. and (5) collecting cases. According to SYSUCC queue case inclusion criteria, all 71 cases meeting inclusion criteria in 2012-2020 of a first affiliated hospital of the university of Zhongshan, all 84 cases meeting inclusion criteria in 2016-2020 of a tumor hospital of the national academy of medical science, and all 55 SFTs meeting inclusion criteria in 2012-2015 of the tumor hospital of the national academy of medical science are screened as three independent verification queues to verify the prediction efficacy of the integrated risk assessment model.
b. Collecting patient clinical pathology information from the electronic medical record systems of the tumor hospitals of the first affiliated hospital of Zhongshan university and the academy of medical science of China, respectively, comprising: the method comprises the steps of determining the sex of a patient, the age of the patient in diagnosis, the type of the specimen, the size of the tumor, the anatomical part of the tumor, follow-up information and the like, wherein the dividing standard of the type of the specimen, the age of the patient in diagnosis, the size of the tumor, the anatomical part of the tumor and the like is the same as SYSUCC queue, the follow-up mode is mainly that the electronic medical record of a medical record system, an imaging examination report and telephone follow-up are inquired, and the main end point of evaluation is PFS.
c. Histopathological evaluation. All HE sections of individual specimens were pathologically diagnosed and reclassified by a soft histopathologist with a high experience, according to WHO classification fifth edition criteria published in month 4 2020. The following histopathological data were collected for each SFTs tissue: nuclear division count, necrosis, nuclear polymorphism, tumor tissue cytopenia, which was evaluated on the same scale as SYSUCC cohort.
Sanger sequencing detects MTOR gene variation in tumor specimens from three independent validation queues SFTs.
IHC detects the expression condition of CD163 and Ki-67 in three independent verification queues, and HALO 2.3 digital pathology quantitative analysis of the CD163 and Ki-67 positive cell density divides the expression condition into high expression and low expression according to the SYSUCC queue CD163 and Ki-67 positive cell density cut-off values.
f. And (3) respectively calculating Nomogram scores of all cases of the three independent verification queues, respectively dividing the cases of the three verification queues into high, medium and low risk groups according to the grouping standard of the SYSUCC queue, and drawing a survival curve by R software.
g. And (3) evaluating the prediction accuracy of the risk assessment model in the discovery queue and three independent verification queues by adopting methods such as C index, ROC Area Under Curve (AUC) and the like, and comparing the prediction accuracy with the 2020WHO classification, the mDemicco model and the G-score model. SYSUCC queue: the C indexes of the integrated risk assessment model, 2020WHO classification, mDemicco model and G-score model are respectively 0.911vs0.787/0.857/0.855, and AUC is respectively 0.921vs 0.790/0.832/0.865; FAHSYSU queue: the C indexes of the integrated risk assessment model, 2020WHO classification, mDemicco model and G-score model are respectively 0.8235 vs0.680/0.760/0.785, and the AUC is respectively 0.85rvs 0.762/0.756/0.731; chcam queue: the C indexes of the integrated risk assessment model, 2020WHO classification, mDemicco model and G-score model are respectively 0.919vs 0.825/0.903/0.806, and AUC is respectively 0.928vs 0.753/0.808/0.822; chcam 2 queue: the C index of the integrated risk assessment model, 2020WHO classification, mDemicco model and G-score model was 0.903vs 0.785/0.814/0.850, respectively, and the AUC was 0.887vs 0.796/0.820/0.698, respectively. I.e., the C-index and AUC of the integrated risk assessment model are superior to 2020WHO classification, mDemicco model and G-score model, and are verified in three verification queues.
In order to more strictly compare the integrated risk assessment model with 2020WHO classification, mDemicco model and G-score model, the present study distinguishes the ending index of each model, taking the ending index reported in the original literature of each model as reference: the integrated risk assessment model and 2020WHO classification employed PFS, no metastasis survival (Metastasis Free Survival, MFS) and no recurrence interval (Recurrence Free Interval, RFI), the mdemico model was used as the ending indicator and the G-score model was used as the ending indicator. The performance of the different risk models was compared by C-index and AUC, scoring according to the scoring criteria of each model. Taking PFS as a ending index, integrating a risk assessment model, 2020WHO classification, mDemicco model and G-score model, wherein the C indexes are respectively 0.893vs 0.768/0.805/0.779, and the AUC is respectively 0.891vs 0.783/0.802/0.797; with MFS as a ending index, integrating a risk assessment model, 2020WHO classification, mDemicco model and G-score model, wherein the C indexes are respectively 0.258 vs. 0.823/0.771/0.736, and the AUC is respectively 0.902 vs. 0.840/0.816/0.778; with RFI as a ending index, the C indexes of the integrated risk assessment model, 2020WHO classification, mDemicco model and G-score model are respectively 0.884vs 0.741/0.773/0.735, and AUC is respectively 0.872vs 0.759/0.785/0.748. I.e. in the case of stricter standards, the performance of the integrated risk assessment model is better.
In an alternative embodiment, as shown in FIG. 2, a risk assessment model is developed in the discovery queue. a, the alignment graph of the risk assessment model contains four variables: nuclear division count, density of Ki-67+ and cd163+ cells, and MTOR variation. It predicts the risk of progression for 3 years, 5 years, 8 years and 10 years per SFT patient. The importance of each variable is ordered according to standard deviation on the nomogram scale. For using the nomograms, a specific point corresponding to each patient is located on each variable axis. The sum of these points is then located on the total point axis and a line is drawn down on the survival axis to determine the probability of PFS for 3, 5, 8 and 10 years. b, scoring and risk layering division standards of each variable in the SFTs risk assessment model. Patients were divided into 3 risk groups according to the Nomogram total score for each tissue specimen: low risk (total score=0), medium risk (0 < total score +.3.37), and high risk (total score > 3.37).
Schematically, an example of SFTs tissue was analyzed and found to have a Ki-67 positive cell density of 1000cells/mm 2 (high expression), MTOR was unchanged, and the CD 163-positive cell density was 100cells/mm 2 (under-expression), the mitosis count is not less than 4/10HPF, and the integrated risk assessment model is used for predicting PFS in 3 years, 5 years, 8 years and 10 years.
The implementation process comprises the following steps: a red line is vertically upwards drawn along the rightmost end of the Ki-67 variable axis to reach the Points axis, a red line is vertically upwards drawn along the leftmost end of the MTOR variable axis to reach the Points axis, a red line is vertically upwards drawn along the leftmost end of the CD163 variable axis to reach the Points axis, a red line is vertically upwards drawn along the rightmost end of the mitosis calculation variable axis to reach the Points axis, a Total score of four variables is 74, a point corresponding to the point is found on the Total Points axis, a straight line is vertically downwards drawn along the point, 3 years, 5 years, 8 years and 10 years of PFS of the SFT patient are respectively 0.812,0.538,0.321 and 0.153. The total score of the four variables of the SFT is 74 points, and the SFT is predicted as a stroke risk group SFTs according to the grouping standard.
In an alternative embodiment, the risk assessment model is equally applicable to central nervous system SFTs, a total of 36 primary cut-edge negative SFTs in the four cohorts are included in the analysis, the integrated risk assessment model is applied to predict 36 primary cut-edge negative central nervous system SFTs prognosis, and compared with 2021WHO central nervous system tumor classification, the analysis results show that the C index (0.870vs 0.781) and AUC (0.802vs 0.712) of the integrated risk assessment model perform better, i.e. the integrated risk assessment model is equally applicable to predict central nervous system SFTs prognosis and is better than 2021WHO central nervous system tumor classification.
In an alternative embodiment, as shown in FIG. 3, a, LASSO regression screens and sequences variables initially include clinical and histopathological factors (including mitotic counts, necrosis, age, sex, tumor size, tumor site, cell number and nuclear polymorphism), immune factors (including CD68+, CD163+, HLA-DPB1+, PD-L1+, CD1+, CD4+, CD11c+, CD20 positive cell density and proliferation index Ki-67 positive cell density) and genetic variation information (including MTOR variation, notch signaling pathway variation (NOTCH 1, NOTCH2, NOTCH3 and CREBBP), ERCC5 variation, TP53 variation, TERT promoter region variation and IDH1 variation). b, LASSO regression cross-validation, fitting and model selection. And c, screening and sequencing the importance of the forest according to the random survival forest VIPM. And d, a variable selected by LASSO regression and random survival forest together.
In an alternative embodiment, as shown in FIG. 4, the integrated risk assessment model is validated in three separate queues. Kaplan-Meier plot shows PFS of SFT patients stratified with risk assessment model: a, SYSUCC queue; b, FAHSYSU queue; c, chcam queue 1; d, chcam queue 2.
In an alternative embodiment, as shown in FIG. 5, the risk assessment model is compared to the world health organization classification and published model. a. b, integrating the c index and ROC curves of the risk assessment model, 2020WHO classification, mDemicco model and G-score model in SYSUCC queue. c. d, the risk assessment model in FAHSYSU queue, 2020WHO classification, mDemicco model, c index of G-score model and ROC curve. e. f, c index and ROC curve of the risk assessment model, WHO classification, mDemicco model, G-score model in CHCAMS queue 1. g. h, c-index and ROC curves for the risk assessment model, 2020WHO classification, mDemicco model and G-score model in CHCAMS cohort 2.
In an alternative embodiment, as shown in FIG. 6, the risk assessment model is compared to 2020WHO classification, mDemicco model and G-score model under more stringent criteria. All NTM primary non-central nervous system SFTs (n=311) that were included in the 4 queues were compared with MFS and RFI as the result index in addition to PFS. and a, respectively adopting PFS, MFS and RFI as ending indexes, integrating a risk assessment model, a WHO classification model, a mDemicco model and a G-score model, and integrating c indexes of SFTs meeting inclusion standards in four queues. b-d, respectively adopting PFS, MFS and RFI as ending indexes, integrating ROC curves of SFTs meeting the inclusion standard in four queues by a risk assessment model, a WHO classification model, a mDemicco model and a G-score model. Kaplan-Meier plots show the PFS of the risk assessment model (e) and 2020WHO classification (f), MFS of the mdemico model (G) and RFI of the G-score model (h) for all primary non-central nervous system NTM patients (n=311).
In an alternative embodiment, as shown in fig. 7, the risk assessment model is presented in primary central nervous system SFTs (NTMs) with tumor margin negative. a, integrating a risk assessment model and a c index of 2020WHO classification. b, integrating a risk assessment model and a ROC curve of 2020WHO classification. c, kaplan-Meier plot shows PFS of primary CNS SFTs with risk stratification via integrated risk assessment model. d, sankey diagram shows 2020WHO classification and intersection of SFT patients risk stratified by integrating risk assessment model.
The above embodiment of the present invention has the following effects by implementing the present invention:
1. compared with 2020WHO classification, mDemicco model and G-score model, the integrated risk assessment model can more accurately predict SFTs prognosis; (the C-index and AUC of the integrated risk assessment model are superior to 2020WHO classification, mDemicco model and G-score model, and are verified in three verification queues.
2. The integrated risk assessment model integrates the genetic variation, the immune infiltration characteristic and the clinical pathology characteristic of the SFTs, and analyzes the possible prognosis factors of the SFTs from a more-dimensional aspect compared with the traditional SFTs prognosis model (only the clinical pathology characteristic is considered).
3. The integrated risk assessment model is also suitable for risk stratification of central nervous system SFTs and is superior to 2021 central nervous system tumor stratification.
While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that changes and modifications may be made without departing from the principles of the invention, such changes and modifications are also intended to be within the scope of the invention.
Claims (8)
1. An SFTs integrated risk assessment system, comprising: the system comprises a feature selection module, a risk assessment module and a progression-free survival assessment module;
the characteristic selection module is used for acquiring selected pathological characteristics, selected immune infiltration characteristics and selected gene variation characteristics corresponding to a patient to be detected;
the risk assessment module is used for inputting the pathological features, the immune infiltration features and the genetic variation features into a risk assessment model so that the risk assessment model determines a risk assessment score according to the pathological features, the immune infiltration features and the genetic variation features;
the non-progress survival assessment module is used for assessing the survival probability of the patient to be detected in each non-progress survival period according to the risk assessment score;
the risk assessment model is obtained by taking selected pathological features, selected immune infiltration features and selected genetic variation features of training samples as inputs and taking risk assessment scores corresponding to the training samples as outputs.
2. The SFTs integrated risk assessment system of claim 1, wherein the selected pathological feature comprises: a mitosis count; the selected immunoinfiltration features include: CD163 positive cell density and Ki-67 positive cell density; the selected genetic variation signature comprises: MTOR gene variation information.
3. The SFTs integrated risk assessment system of claim 1, further comprising:
and constructing an alignment chart corresponding to the risk assessment model according to the selected pathological features, the selected immune infiltration features, the selected genetic variation features, the risk assessment score and the probability of each progression-free survival period.
4. The SFTs integrated risk assessment system of claim 1, wherein in training the risk assessment model, selected pathological features, selected immune infiltration features, and selected genetic variation features of a training sample are determined by:
acquiring an SFTs clinical pathological feature set, an SFTs gene variation feature set and an SFTs immune infiltration feature set;
according to a preset LASSO regression model, a first training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
according to a preset random survival forest model, a second training subset is screened from the SFTs clinical pathology feature set, the SFTs gene variation feature set and the SFTs immune infiltration feature set;
and constructing an intersection of the first training subset and the second training subset, and determining selected pathological features, selected immune infiltration features and selected genetic variation features of the training sample according to the intersection.
5. The SFTs integrated risk assessment system of claim 4, wherein the SFTs clinical pathology feature set comprises: tumor site, patient age, patient sex, tumor size, nuclear division count, necrosis, nuclear polymorphism, tumor tissue cytopenia, patient radiation information, and patient chemotherapy information.
6. The SFTs integration risk assessment system of claim 4, wherein the SFTs genetic variation feature set comprises: MTOR variation, NOTCH1 signaling pathway variation, ERCC5 variation, TP53 variation, TERT promoter region variation, and IDH1 variation.
7. The SFTs integrated risk assessment system of claim 4 wherein the SFTs immunoinfiltration feature set comprises: densities of PD-L1, ki-67, CD68, CD163, HLA-DPB1, CD3, CD4, CD8, FOXP3, CD11c and CD20 positive cells.
8. The SFTs integrated risk assessment system of claim 1, further comprising: a risk division module;
the risk classification module is used for classifying the patients to be detected into a low risk group, a medium risk group or a high risk group according to the risk assessment score;
wherein when the risk assessment score is less than or equal to a first threshold, dividing the patient to be detected into a low risk group; dividing the patient to be detected into risk groups when the risk assessment score is greater than a first threshold and less than or equal to a second threshold; dividing the patient to be detected into high-risk groups when the risk assessment score is greater than a second threshold value; the second threshold is greater than the first threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311186054.7A CN117174323A (en) | 2023-09-14 | 2023-09-14 | SFTs integration risk assessment system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311186054.7A CN117174323A (en) | 2023-09-14 | 2023-09-14 | SFTs integration risk assessment system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117174323A true CN117174323A (en) | 2023-12-05 |
Family
ID=88942767
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311186054.7A Pending CN117174323A (en) | 2023-09-14 | 2023-09-14 | SFTs integration risk assessment system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117174323A (en) |
-
2023
- 2023-09-14 CN CN202311186054.7A patent/CN117174323A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cao et al. | Validation of American college of radiology ovarian-adnexal reporting and data system ultrasound (O-RADS US): analysis on 1054 adnexal masses | |
| US20220291219A1 (en) | Systems and compositions for diagnosing barrett's esophagus and methods of using the same | |
| Ueno et al. | New criteria for histologic grading of colorectal cancer | |
| US20160341731A1 (en) | Method and system for classification and quantitative analysis of cell types in microscopy images | |
| CN101689220A (en) | The system and method that be used for the treatment of, diagnosis and prospective medicine illness takes place | |
| Wang et al. | Computed-tomography-based radiomics model for predicting the malignant potential of gastrointestinal stromal tumors preoperatively: a multi-classifier and multicenter study | |
| CN113270188A (en) | Method and device for constructing prognosis prediction model of patient after esophageal squamous carcinoma radical treatment | |
| CN114891887A (en) | Method for screening triple negative breast cancer prognosis gene marker | |
| Choe et al. | CT radiomics-based prediction of anaplastic lymphoma kinase and epidermal growth factor receptor mutations in lung adenocarcinoma | |
| Ameye et al. | A scoring system to differentiate malignant from benign masses in specific ultrasound‐based subgroups of adnexal tumors | |
| Sethi et al. | Patient and tumor factors at diagnosis in a multi‐ethnic primary head and neck squamous cell carcinoma cohort | |
| Gao et al. | A radiogenomics biomarker based on immunological heterogeneity for non-invasive prognosis of renal clear cell carcinoma | |
| Wen et al. | Deep learning in digital pathology for personalized treatment plans of cancer patients | |
| Zhang et al. | Radiomics based on CECT in differentiating kimura disease from lymph node metastases in head and neck: a non-invasive and reliable method | |
| Segnan et al. | Estimate of false-positive breast cancer diagnoses from accuracy studies: a systematic review | |
| CN117174323A (en) | SFTs integration risk assessment system | |
| Xu et al. | Prediction breast molecular typing of invasive ductal carcinoma based on dynamic contrast enhancement magnetic resonance imaging radiomics characteristics: a feasibility study | |
| CN115602327A (en) | Construction method of prediction model for lung nodule lung cancer risk | |
| Guo et al. | Evaluation of the peritumoral features using radiomics and deep learning technology in non-spiculated and noncalcified masses of the breast on mammography | |
| Guo et al. | Machine learning for identifying benign and malignant of thyroid tumors: A retrospective study of 2,423 patients | |
| CN112259231A (en) | High-risk gastrointestinal stromal tumor patient postoperative recurrence risk assessment method and system | |
| Barrenada et al. | The ADNEX risk prediction model for ovarian cancer diagnosis: A systematic review and meta-analysis of external validation studies | |
| Duran-Ranada et al. | Diagnostic accuracy performance of the international ovarian tumour analysis (IOTA) simple rules and assessment of different neoplasias in the adnexa (ADNEX) model for identifying benign or malignant adnexal masses against histopathological diagnosis in p | |
| Guo et al. | Machine learning for identifying benign and malignant of thyroid tumors: A retrospective study of, patients | |
| Abdalla et al. | Can Replacing CA125 with HE4 in Risk of Malignancy Indices 1–4 Improve Diagnostic Performance in the Presurgical Assessment of Adnexal Tumors? |
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 |