CN113160979A - Machine learning-based liver cancer patient clinical prognosis prediction method - Google Patents
Machine learning-based liver cancer patient clinical prognosis prediction method Download PDFInfo
- Publication number
- CN113160979A CN113160979A CN202011505660.7A CN202011505660A CN113160979A CN 113160979 A CN113160979 A CN 113160979A CN 202011505660 A CN202011505660 A CN 202011505660A CN 113160979 A CN113160979 A CN 113160979A
- Authority
- CN
- China
- Prior art keywords
- features
- sample data
- prediction
- survival
- liver cancer
- 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
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
Abstract
The invention discloses a machine learning-based liver cancer patient clinical prognosis prediction method, which comprises the following steps: collecting sample data of the corresponding hepatocellular carcinoma according to the predetermined prediction characteristics; sorting the prediction characteristics according to the sample data, and selecting partial prediction characteristics according to a sorting result; determining candidate predictive features related to overall survival rate from the selected predictive features by using a proportional risk regression model; and selecting 20 optimal prediction features from the candidate prediction features by using a gradient enhanced survival classifier as survival prediction factors to construct a liver cancer patient prognosis survival model. The method can effectively identify patients with high death risk related to HCC.
Description
Technical Field
The invention relates to a liver cancer survival analysis technology, in particular to a machine learning-based prediction method for clinical prognosis of a liver cancer patient.
Background
Liver cancer accounts for 8.2% of all cancer deaths worldwide. The major risk factors for hepatocellular carcinoma (HCC) include Hepatitis B Virus (HBV) or Hepatitis C Virus (HCV) infection and alcoholic or non-alcoholic liver disease. The clinical efficacy of liver cancer patients depends to a large extent on tumor burden, treatment modality and liver function. A number of staging systems and predictive/prognostic models are developed to assess liver function reserve.
To date, several staging systems have been used to predict the clinical efficacy of different treatments for liver cancer patients. In addition, liver function parameters, including the Child-Pugh classification system, albumin and bilirubin levels, the International Normalized Ratio (INR) and alkaline phosphatase (ALP), and platelet rates, have also been shown to be independent prognostic factors affecting the Overall Survival (OS) of HCC patients. However, there is no general consensus on the best prognostic models, and there is a need for new methods for predicting clinical prognosis and recurrence in liver cancer patients.
Disclosure of Invention
The invention mainly aims to provide a machine learning-based method for predicting the clinical prognosis of a liver cancer patient, so as to solve the problems in the prior art.
The method for predicting the clinical prognosis of the liver cancer patient based on machine learning provided by the embodiment of the invention comprises the following steps: collecting sample data of the corresponding hepatocellular carcinoma according to the predetermined prediction characteristics; sorting the prediction characteristics according to the sample data, and selecting partial prediction characteristics according to a sorting result; determining candidate predictive features related to overall survival rate from the selected predictive features by using a proportional risk regression model; and selecting 20 optimal prediction features from the candidate prediction features by using a gradient enhanced survival classifier as survival prediction factors to construct a liver cancer patient prognosis survival model.
Wherein the predetermined types of predictive features include: clinical features, laboratory examination features, and flow-through testing features.
Wherein the 20 optimal predictive features include: 1 clinical feature: tumor size; 14 laboratory examination features: CRP, γ -GGT, ALP, N/L, GRAN, TBTL, BUN, Cr,% EO, PTA, HGB, RBC, CHE,% Lymph; 5 flow assay features: CD8+ PD-1+ TIGIT + TIM-3+, CD4+ TIGIT + TIM-3+, CD4+ PD-1+ TIGIT + TIM-3+, CD8+ Tcm, CD8+ Tem.
Wherein, the prediction characteristics of the top 30%, 40% or 50% of the ranking are selected according to the ranking result.
Wherein the step of ordering the predicted features according to the sample data comprises: calculating the sample data by using multiple modes, and sequencing the characteristics corresponding to the sample data according to the calculation result, wherein at least one of the following modes is adopted for calculating the sample data: ANOVA F test, mutual information, Pearson correlation test, Spearman correlation test, Kendall tau rank correlation coefficient, and logistic regression.
Wherein, prior to the step of ordering the predicted features according to the sample data, the method further comprises: preprocessing the predicted features, specifically comprising: and adding the pseudo count 1 into the sample data, then carrying out log2 transformation, deleting the prediction characteristic lacking the sample data, and deleting the prediction characteristic of the same sample data.
Wherein the step of using a proportional risk regression model to determine candidate predictive features associated with overall survival from the selected predictive features comprises: the 48 candidate predictive features associated with overall survival were determined using a proportional Risk regression model to remove predictive features with a consistency index greater than 0.5 and a log likelihood ratio test P-value less than or equal to 0.1.
Wherein the method further comprises: and fitting the sample data by using a recursive characteristic elimination method of cross validation.
Wherein the method further comprises: obtaining hepatocellular carcinoma sample data, randomly dividing the sample data into a training set and a verification set, and constructing the prognosis survival model of the liver cancer patient according to the sample data of the training set.
Wherein the method further comprises: and performing 13 times of cross validation on the training set by using the GBS model according to the selected 20 optimal prediction characteristics to construct the prognosis survival model of the liver cancer patient.
According to the technical scheme of the invention, a plurality of features are selected from the predetermined feature types, and a machine learning-based method is adopted to construct the prognosis survival model of the liver cancer patient, statistics shows that the risk score generated by the GBS model shows a higher consistency index (c-index) in the training set and the verification set 1 and set 2. The GBS classifier can classify patients into high, medium, and low risk subgroups based on death in a dataset. In addition, higher risk scores are positively correlated with higher clinical staging and portal vein cancer emboli (PVTT). Furthermore, subgroup analysis on Child-Pugh stratification, Barcelona Clinical Liver Cancer (BCLC) staging and PVTT status supports prognostic relevance for GBS-based risk algorithm prediction independent of traditional tumor staging systems.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a flow chart of a method for predicting the clinical prognosis of a patient with liver cancer based on machine learning according to an embodiment of the present invention;
FIG. 2 is a diagram of 48 characteristic Overall Survival (OS) related candidate variables of a liver cancer patient according to an embodiment of the present invention;
FIGS. 3A and 3B are schematic diagrams of prognostic multi-parameter risk scores for liver cancer patients, according to embodiments of the present invention;
FIG. 4 is a schematic diagram of 20 optimal predictive features according to an embodiment of the invention;
FIGS. 5A-5F are schematic diagrams of OS KM survival analysis curves for prognosis of liver cancer patients according to embodiments of the present invention;
FIGS. 6A and 6B are schematic diagrams of time-dependent ROC analysis and risk scoring based on machine learning and other conventional hierarchical pattern analysis for liver cancer patients according to embodiments of the present invention;
FIGS. 7A-7H are schematic diagrams of prognostic RFS KM survival analysis curves for liver cancer patients, according to embodiments of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical solutions provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
According to the embodiment of the invention, the method for predicting the clinical prognosis of the liver cancer patient based on machine learning is provided, a certain amount of sample data of the HCC patient is collected firstly, and is randomly divided into a training set and a verification set (set 1), and in addition, other sample data of the HCC patient, which is collected independently, can be used as an independent verification set (set 2). A machine-learned survival model can be constructed based on training set data, and with reference to fig. 1, the method includes the following steps:
step S102, collecting sample data of corresponding hepatocellular carcinoma according to predetermined prediction characteristics;
in the present application, the types of predicted features may include: clinical features, laboratory examination features, and flow-through testing features. Referring to Table 1, conventional clinical characteristics may include tumor size, portal vein cancer emboli (PVTT), Child-Pugh classification, BCLC classification, and the like.
TABLE 1 clinical characteristics of hepatocellular carcinoma patients
Referring to Table 2, the laboratory test features (otherwise known as laboratory parameters) may include γ -GGT, CRP, ALP, HBV-DNA, AFP levels, and the like.
TABLE 2 laboratory characterization of hepatocellular carcinoma patients
Peripheral blood was collected from all patients prior to treatment, Peripheral Blood Mononuclear Cells (PBMCs) were isolated, and multi-color flow cytometry assays were performed. Immune checkpoint molecules for surface staining of T cells and antibodies used for staining are as follows PD-1-BV711 (clone EH12.1, BD #564017), TIGIT-PE-CY 7 (clone MBSA43, eBioscience #25-9500-42), TIM-3-BV650 (clone 7D3, BD #565564), LAG-3-APC (clone 3DS223H, eBioscience #17-2239-42), and CTLA-4 (clone BNI3, BD # 555853). Sub-populations of CD4 and CD 8T cells
T (Tn) cell population (CD45RA + CCR7+), central memory T (Tcm) cell population (CD45RA-CCR7+), effector memory T (Tem) cell population (CD45RA-CCR7-) and terminally differentiated T (Temra) cell population (CD45RA + CCR 7-).
Data collection can be performed using an LSR Fortessa flow cytometer (BD Biosciences), data processing and analysis can be performed using FlowJo software, and flow cytometer detection indices for hepatocellular carcinoma patients are shown in table 3.
TABLE 3 flow cytometry detection indices for hepatocellular carcinoma patients
And S104, sorting the prediction features according to the sample data, and selecting partial prediction features according to a sorting result.
Before this step, the predicted features may be preprocessed, specifically, three types of features are stacked into one integrated machine learning framework, in order to reduce outliers and make numerical features approximate to gaussian distribution, the values of all features are log2 transformed after adding pseudo count 1; features lacking values above 20% are deleted from the sample, and features with the same values above 97% are deleted. Through the above processing, the low variance characteristic of low information content and high background noise is removed.
When the predicted features are ranked, the sample data can be calculated in various modes, and the features corresponding to the sample data are ranked according to the calculation result. The following six unsupervised methods can be used simultaneously or one of them can be used to calculate the sample data: ANOVA F test, mutual information, Pearson correlation test, Spearman correlation test, Kendall tau rank correlation coefficient, and logistic regression. Then, the prediction features of the top 30%, 40% or 50% of the ranking are selected according to the ranking result, and preferably the prediction features of the top 40% of the ranking can be selected for subsequent processing.
And step S106, determining candidate prediction characteristics related to the overall survival rate from the selected prediction characteristics by using a proportional risk regression model.
After data preprocessing, a proportional Risk regression (CoxPH) model is used to remove features that do not meet the threshold criteria, such as (concordance index [ C-index ] > 0.5 and features with log likelihood ratio test P-value ≦ 0.1, and 48 features are determined as candidate predicted features for OS correlation, see FIG. 2.
And S108, selecting 20 optimal prediction features from the candidate prediction features by using a gradient enhanced survival classifier as survival prediction factors to construct a liver cancer patient prognosis survival model.
Data were fitted using cross-validated Recursive Feature Elimination (RFECV) and the highest accuracy feature was selected. In order to obtain consistent features, the hierarchical shuffle-split cross-validation iterator was subjected to 20 iterations and 10 split iterations, with test sizes ranging from 0.15 to 0.34 and step sizes of 0.01.
In the final model, 20 optimal features are selected by using a gradient enhanced survival (GBS) classifier to construct a final survival predictor, and referring to fig. 3A and 3B, it is shown that tumor and immunobiological characteristics can greatly promote the prediction of patient prognosis in the peripheral system. Referring to fig. 4, the selected 20 optimal features include:
1 general Clinical signature (Clinical _ Index): tumor size (diameter (cm));
14 Laboratory parameters (Laboratory _ Index): CRP, γ -GGT, ALP, N/L,% GRAN, TBTL, BUN, Cr,% EO, PTA, HGB, RBC, CHE,% Lymph;
5T cell functional parameters (Immune _ Index): CD8+ PD-1+ TIGIT + TIM-3+, CD4+ TIGIT + TIM-3+, CD4+ PD-1+ TIGIT + TIM-3+, CD8+ Tcm, CD8+ Tem. Wherein, the T cell function parameter can also be called as flow test characteristic or flow cytometry detection index.
Based on the selected 20 features, the training set was modeled with 13 cross-validation (10 segmentation and stratification transformations and 3 segmentation and stratification k-fold) using the GBS model. In each fold, the hyper-parametric optimization was adjusted by performing an exhaustive cross-validation grid search using 0.3 of the training set as the test set, and then re-fitting the model into the training set using the best scoring parameters. The performance of our survival predictor was evaluated using a truncated c-index (c-static) and the area under the average cumulative/dynamic curve (AUC).
Referring to table 4, table 4 shows the correlation between the risk scores generated by the 3-queue GBS models of training trials, certificate sets set 1, set 2, etc. and OS (months) according to an embodiment of the present invention. The independent validation sets set 1 and set 2 were predicted using a training set program. The relationship between the risk score generated by the GBS model and OS (months) was examined in three data sets. There is a significant linear correlation between the two indices (Pearson correlation, P-value < 0.0001). Although median follow-up times for surviving patients vary widely among the three cohorts, from 12.4 months to 19.3 months, the predictive risk scoring system showed higher concordance indices (c-index) in the training and validation sets set 1, set 2 of 0.844, 0.827 and 0.806, respectively.
TABLE 4
Data analysis may use version 3.6 of the R statistical software, using the chi-square test for categorical data and the Wilcoxon rank-sum test for continuous variables. OS was assessed using Kaplan-Meier (KM) survival analysis. For the classification of both subgroups, the risk score is 0 as a cutoff value. For the three subgroup classifications, the cut-off values for the high, medium and low risk scores were determined using the X-tile plots of the training set. Time-dependent Receiver Operating Characteristic (ROC) curve analysis was used to analyze the prediction and prognostic accuracy of each feature.
Referring to fig. 5A-5F, Kaplan-Meier curves show that the machine learning classifier can effectively classify patients into subgroups with higher risk of death (Group _ High) or lower subgroups (Group _ Low). Whether two (fig. 5A-5C) or three (fig. 5D-5F), the OS of the low-risk and/or medium-risk subgroups was significantly improved over the high-risk subgroup in all three datasets (all comparisons P < 0.05). For example, referring to FIG. 5B, in the two subfraction class of validation set 1, the mortality rate for patients with Low risk score (Low) was significantly lower than for patients with High risk score (High) (risk ratio [ HR ]:0.12, 95% confidence interval [ CI ]:0.04-0.35, P < 0.0001). Also, similar results were found in the verification set 2.
Referring to fig. 6A and 6B, the present application determined the prognosis and prediction accuracy of the GBS risk score classifier by performing a time-dependent ROC analysis, and the results showed that the area under the curve (AUC) of the risk score was very high in both validation sets (0.886, 0.858, 0.917 at validation set 1: 6, 12, 18 months, respectively; 0.924, 0.868, 0.884 at validation set 2: 6, 12, 18 months, respectively). Analysis of the univariate Cox model for multiple prognostic indicators shows that the stratification potential of GBS-derived risk scores is superior to other traditional models.
Referring to FIGS. 7A-7H, Kaplan-Meier survival analysis was performed for recurrence-free survival (RFS) time for all three data sets. Analysis confirmed that in all cohorts, the low risk score subgroup had longer RFS times than the high risk score subgroup (P < 0.05 for all comparisons). Patients were divided into three groups of low, medium and high risk with similar results (P < 0.05 for all comparisons).
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The above description is only an example of the present invention, and is not intended to limit the present invention, and it is obvious to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims (10)
1. A method for predicting the clinical prognosis of a liver cancer patient based on machine learning is characterized by comprising the following steps:
collecting sample data of the corresponding hepatocellular carcinoma according to the predetermined prediction characteristics;
sorting the prediction characteristics according to the sample data, and selecting partial prediction characteristics according to a sorting result;
determining candidate predictive features related to overall survival rate from the selected predictive features by using a proportional risk regression model;
and selecting 20 optimal prediction features from the candidate prediction features by using a gradient enhanced survival classifier as survival prediction factors to construct a liver cancer patient prognosis survival model.
2. The method of claim 1, wherein the predetermined type of predicted feature comprises: clinical features, laboratory examination features, and flow-through testing features.
3. The method of claim 2, wherein the 20 optimal predictive features comprise:
1 clinical feature: tumor size;
14 laboratory examination features: CRP, γ -GGT, ALP, N/L, GRAN, TBTL, BUN, Cr,% EO, PTA, HGB, RBC, CHE,% Lymph;
5 flow assay features: CD8+ PD-1+ TIGIT + TIM-3+, CD4+ TIGIT + TIM-3+, CD4+ PD-1+ TIGIT + TIM-3+, CD8+ Tcm, CD8+ Tem.
4. The method of claim 1, wherein the top 30%, 40%, or 50% of the predicted features are selected according to the ranking result.
5. The method of claim 1, wherein said step of ordering predicted features according to said sample data comprises:
calculating the sample data by using multiple modes, and sequencing the characteristics corresponding to the sample data according to the calculation result, wherein at least one of the following modes is adopted for calculating the sample data: ANOVA F test, mutual information, Pearson correlation test, Spearman correlation test, Kendall tau rank correlation coefficient, and logistic regression.
6. The method according to claim 1 or 5, wherein prior to the step of ordering predicted features according to the sample data, the method further comprises:
preprocessing the predicted features, specifically comprising: and adding the pseudo count 1 into the sample data, then carrying out log2 transformation, deleting the prediction characteristic lacking the sample data, and deleting the prediction characteristic of the same sample data.
7. The method of claim 6, wherein the step of using a proportional hazards regression model to determine candidate predictive features from the selected predictive features that correlate to overall survival comprises:
the 48 candidate predictive features associated with overall survival were determined using a proportional Risk regression model to remove predictive features with a consistency index greater than 0.5 and a log likelihood ratio test P-value less than or equal to 0.1.
8. The method of claim 7, further comprising:
and fitting the sample data by using a recursive characteristic elimination method of cross validation.
9. The method of claim 1, further comprising:
obtaining hepatocellular carcinoma sample data, randomly dividing the sample data into a training set and a verification set, and constructing the prognosis survival model of the liver cancer patient according to the sample data of the training set.
10. The method of claim 9, further comprising:
and performing 13 times of cross validation on the training set by using the GBS model according to the selected 20 optimal prediction characteristics to construct the prognosis survival model of the liver cancer patient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011505660.7A CN113160979A (en) | 2020-12-18 | 2020-12-18 | Machine learning-based liver cancer patient clinical prognosis prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011505660.7A CN113160979A (en) | 2020-12-18 | 2020-12-18 | Machine learning-based liver cancer patient clinical prognosis prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113160979A true CN113160979A (en) | 2021-07-23 |
Family
ID=76882632
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011505660.7A Pending CN113160979A (en) | 2020-12-18 | 2020-12-18 | Machine learning-based liver cancer patient clinical prognosis prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113160979A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116564421A (en) * | 2023-06-08 | 2023-08-08 | 苏州卫生职业技术学院 | Method for constructing prognosis model related to copper death of acute myelogenous leukemia patient |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107657149A (en) * | 2017-09-12 | 2018-02-02 | 中国人民解放军军事医学科学院生物医学分析中心 | System for predicting liver cancer patient prognosis |
| CN110577998A (en) * | 2019-01-31 | 2019-12-17 | 上海交通大学医学院附属仁济医院 | Construction of molecular model for predicting postoperative early recurrence risk of liver cancer and application evaluation thereof |
| CN111094977A (en) * | 2017-07-13 | 2020-05-01 | 古斯塔夫·鲁西研究所 | Imaging tools based on imaging omics to monitor tumor lymphocyte infiltration and prognosis in anti-PD-1/PD-L1 treated tumor patients |
| CN111122865A (en) * | 2019-12-12 | 2020-05-08 | 中山大学 | Marker for liver cancer prognosis prediction based on CD11b and CD169 protein molecules |
| CN111724903A (en) * | 2020-06-29 | 2020-09-29 | 北京市肿瘤防治研究所 | System for predicting gastric cancer prognosis in a subject |
| WO2020212586A1 (en) * | 2019-04-18 | 2020-10-22 | INSERM (Institut National de la Santé et de la Recherche Médicale) | Methods for the treatment and prognosis of cancer |
| CN112011616A (en) * | 2020-09-02 | 2020-12-01 | 复旦大学附属中山医院 | Immune gene prognosis model for predicting hepatocellular carcinoma tumor immune infiltration and postoperative survival time |
-
2020
- 2020-12-18 CN CN202011505660.7A patent/CN113160979A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111094977A (en) * | 2017-07-13 | 2020-05-01 | 古斯塔夫·鲁西研究所 | Imaging tools based on imaging omics to monitor tumor lymphocyte infiltration and prognosis in anti-PD-1/PD-L1 treated tumor patients |
| CN107657149A (en) * | 2017-09-12 | 2018-02-02 | 中国人民解放军军事医学科学院生物医学分析中心 | System for predicting liver cancer patient prognosis |
| CN110577998A (en) * | 2019-01-31 | 2019-12-17 | 上海交通大学医学院附属仁济医院 | Construction of molecular model for predicting postoperative early recurrence risk of liver cancer and application evaluation thereof |
| WO2020212586A1 (en) * | 2019-04-18 | 2020-10-22 | INSERM (Institut National de la Santé et de la Recherche Médicale) | Methods for the treatment and prognosis of cancer |
| CN111122865A (en) * | 2019-12-12 | 2020-05-08 | 中山大学 | Marker for liver cancer prognosis prediction based on CD11b and CD169 protein molecules |
| CN111724903A (en) * | 2020-06-29 | 2020-09-29 | 北京市肿瘤防治研究所 | System for predicting gastric cancer prognosis in a subject |
| CN112011616A (en) * | 2020-09-02 | 2020-12-01 | 复旦大学附属中山医院 | Immune gene prognosis model for predicting hepatocellular carcinoma tumor immune infiltration and postoperative survival time |
Non-Patent Citations (3)
| Title |
|---|
| 廖锡文: "乙肝相关性原发性肝细胞癌切除术后行TACE致ALT,AST变化相关的单核苷酸多态性特征筛查及验证", 《中国优秀博硕士学位论文全文数据库(博士)医药卫生科技辑》 * |
| 李琳: "基于机器学习方法的肝癌预后预测模型研究", 《中国优秀博硕士学位论文全文数据库(硕士)医药卫生科技辑》 * |
| 高世明，李旭，郭晓东: "《现代医院诊疗常规》", 30 September 2002, 安徽科学技术出版社 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116564421A (en) * | 2023-06-08 | 2023-08-08 | 苏州卫生职业技术学院 | Method for constructing prognosis model related to copper death of acute myelogenous leukemia patient |
| CN116564421B (en) * | 2023-06-08 | 2024-01-30 | 苏州卫生职业技术学院 | Method for constructing prognosis model related to copper death of acute myelogenous leukemia patient |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TW200828053A (en) | A method for grid-based data clustering | |
| CN110634563A (en) | Differential diagnosis device for diabetic nephropathy and non-diabetic nephropathy | |
| Call et al. | Incidence of chronic lymphocytic leukemia and high‐count monoclonal B‐cell lymphocytosis using the 2008 guidelines | |
| CN109284626A (en) | Random forests algorithm towards difference secret protection | |
| CN112185549A (en) | Esophageal squamous carcinoma risk prediction method based on clinical phenotype and logistic regression analysis | |
| CN111128385B (en) | Prognosis early warning system for esophageal squamous carcinoma and application thereof | |
| CN111161882A (en) | Breast cancer life prediction method based on deep neural network | |
| CN111062425B (en) | Unbalanced data set processing method based on C-K-SMOTE algorithm | |
| CN112017791B (en) | System for determining prognosis condition of liver cancer patient based on artificial neural network model | |
| Montanari et al. | Recursive partitioning analysis of prognostic factors in post‐transplant lymphoproliferative disorders (PTLD): a 120 case single institution series | |
| CN111321225B (en) | Application of lncRNA evaluation system in prediction of tumor immunotherapy effect | |
| CN114373544A (en) | Method, system and device for predicting membranous nephropathy based on machine learning | |
| CN113215254A (en) | Immune-clinical characteristic combined prediction model for evaluating lung adenocarcinoma prognosis | |
| Meiseles et al. | Explainable machine learning for chronic lymphocytic leukemia treatment prediction using only inexpensive tests | |
| CN113160979A (en) | Machine learning-based liver cancer patient clinical prognosis prediction method | |
| Roca et al. | A cross entropy test allows quantitative statistical comparison of t-SNE and UMAP representations | |
| CN113764100B (en) | Prediction model for immunotherapy super-progression risk of liver cancer patient, construction method and application thereof | |
| Ramos et al. | An interpretable approach for lung cancer prediction and subtype classification using gene expression | |
| Cosma et al. | Prostate cancer: early detection and assessing clinical risk using deep machine learning of high dimensional peripheral blood flow cytometric phenotyping data | |
| Casey et al. | A machine learning approach to prostate cancer risk classification through use of RNA sequencing data | |
| CN114898874A (en) | Prognosis prediction method and system for renal clear cell carcinoma patient | |
| CN110442798B (en) | Spam comment user group detection method based on network representation learning | |
| CN113782087A (en) | Chronic lymphocytic leukemia SSCR risk model and establishment method and application thereof | |
| CN113584175A (en) | Group of molecular markers for evaluating renal papillary cell carcinoma progression risk and screening method and application thereof | |
| CN116206681A (en) | Method for evaluating prognostic gene pair value of immune infiltration cell 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210723 |
|
| RJ01 | Rejection of invention patent application after publication |