CN109599181B - Survival prediction system and prediction method for T3-LARC patient before treatment - Google Patents
Survival prediction system and prediction method for T3-LARC patient before treatment Download PDFInfo
- Publication number
- CN109599181B CN109599181B CN201910017582.7A CN201910017582A CN109599181B CN 109599181 B CN109599181 B CN 109599181B CN 201910017582 A CN201910017582 A CN 201910017582A CN 109599181 B CN109599181 B CN 109599181B
- Authority
- CN
- China
- Prior art keywords
- data
- module
- survival
- risk factor
- factor
- 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.)
- Active
Links
- 230000004083 survival effect Effects 0.000 title claims abstract description 134
- 238000011282 treatment Methods 0.000 title claims abstract description 29
- 238000000034 method Methods 0.000 title claims description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 45
- 238000000556 factor analysis Methods 0.000 claims abstract description 25
- 206010028980 Neoplasm Diseases 0.000 claims description 46
- 238000004140 cleaning Methods 0.000 claims description 31
- 102000012406 Carcinoembryonic Antigen Human genes 0.000 claims description 15
- 108010022366 Carcinoembryonic Antigen Proteins 0.000 claims description 15
- 210000001165 lymph node Anatomy 0.000 claims description 15
- 238000002203 pretreatment Methods 0.000 claims description 15
- 238000012216 screening Methods 0.000 claims description 15
- 210000000436 anus Anatomy 0.000 claims description 11
- 230000009545 invasion Effects 0.000 claims description 11
- 210000002966 serum Anatomy 0.000 claims description 9
- 230000004069 differentiation Effects 0.000 claims description 8
- 210000003195 fascia Anatomy 0.000 claims description 7
- 210000000713 mesentery Anatomy 0.000 claims description 7
- 201000010879 mucinous adenocarcinoma Diseases 0.000 claims description 6
- 210000003097 mucus Anatomy 0.000 claims description 6
- 208000009956 adenocarcinoma Diseases 0.000 claims description 5
- 238000009098 adjuvant therapy Methods 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 230000009790 vascular invasion Effects 0.000 claims description 4
- 206010027476 Metastases Diseases 0.000 claims description 3
- 210000004204 blood vessel Anatomy 0.000 claims description 3
- 230000009401 metastasis Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims 1
- 238000004393 prognosis Methods 0.000 abstract description 6
- 238000011337 individualized treatment Methods 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 29
- 238000012549 training Methods 0.000 description 24
- 230000006870 function Effects 0.000 description 9
- 108010014172 Factor V Proteins 0.000 description 6
- 238000005070 sampling Methods 0.000 description 6
- 238000012795 verification Methods 0.000 description 6
- 201000010099 disease Diseases 0.000 description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 5
- 238000003384 imaging method Methods 0.000 description 5
- 208000015634 Rectal Neoplasms Diseases 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 206010038038 rectal cancer Diseases 0.000 description 4
- 201000001275 rectum cancer Diseases 0.000 description 4
- 230000005856 abnormality Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000008676 import Effects 0.000 description 3
- 230000002980 postoperative effect Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000011256 aggressive treatment Methods 0.000 description 2
- 238000002512 chemotherapy Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000009099 neoadjuvant therapy Methods 0.000 description 2
- 230000007170 pathology Effects 0.000 description 2
- 238000001959 radiotherapy Methods 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000009885 systemic effect Effects 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 206010061819 Disease recurrence Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001574 biopsy Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 238000000551 statistical hypothesis test Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a survival prediction system for a T3-LARC patient before treatment, which comprises a risk factor acquisition module, a risk factor preprocessing module, a risk factor comprehensive analysis module, an operator selection module, a survival model generation module and a survival rate display module. The invention can provide individualized risk factor analysis of patients for clinical doctors before treatment, predict the recurrence rate and death rate of the patients in N years after operation, provide reference basis for the clinicians to formulate individualized treatment and follow-up schemes, have important clinical significance, and can greatly improve the prognosis of the patients, prolong the life cycle and improve the quality of life.
Description
Technical Field
The invention relates to the technical field of survival prediction, in particular to a survival prediction system and a survival prediction method for a T3-LARC patient before treatment.
Background
The patients with locally advanced rectal cancer (T3-LARC) at T3 are the most common type with the largest prognosis difference in the first rectal cancer at present, and individualized survival prediction is carried out on the patients at the early stage of treatment, so that early treatment decision making is facilitated, and the improvement on the overall prognosis is greatly facilitated. The related research at home and abroad only analyzes the prognostic influence of 1-2 specific imaging factors and has no selectivity on the stage of a patient, so that model software special for individual imaging of T3-LARC specific population for comprehensively evaluating and predicting survival risk cannot be obtained. At present, only a few researches adopt postoperative pathological risk factors to construct a rectal cancer survival prediction model, and the comprehensive analysis and survival model construction of the risk factors based on the imaging characteristics before treatment for a T3-LARC patient are not available. In summary, the reasons for the non-specificity of patient selection (not for T3-LARC), the incompleteness of the factor analysis (optionally investigating certain 1-2 factors), and the hysteresis of the prediction (prediction based on post-operative pathology is already at the end of treatment, whereas preliminary imaging can achieve prediction before treatment) in the current technical background may be: the conventional research has the defects of small sample size and unreasonable design, and only meets the significance of a certain angle for discussing a certain factor or the prediction of postoperative pathology, but cannot discuss a probability prediction model for combining multiple factors before treatment from a comprehensive angle.
In addition, the prior literature is directed to studying the influence of a certain MRI sign on the prognosis, and does not comprehensively analyze and obtain a prediction model.
Disclosure of Invention
In view of the above technical problems in the related art, the present invention provides a survival prediction system and method for T3-LARC patients before treatment, which can overcome the above disadvantages of the prior art.
In order to achieve the technical purpose, the technical scheme of the invention is realized as follows:
a survival prediction system aiming at a T3-LARC patient before treatment comprises a risk factor acquisition module, a risk factor preprocessing module, a risk factor comprehensive analysis module, an operator selection module, a survival model generation module and a survival rate display module, wherein,
the risk factor acquisition module is used for acquiring data of a T3-LARC patient;
the risk factor preprocessing module is used for preprocessing the data acquired by the risk factor acquisition module to obtain meaningful risk factors in single factor analysis;
the risk factor comprehensive analysis module is used for further performing multi-factor survival analysis on meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain risk factors capable of independently influencing the survival of the patient;
the operator selection module is used for the clinician to select the content required to be predicted;
the survival model generation module is used for receiving the information transmitted by the risk factor comprehensive analysis module and the operator selection module at the same time, generating a corresponding survival model editing program by using R-software, and finally outputting a corresponding survival model;
the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirements of the clinician.
Further, the data collected by the risk factor collecting module comprises the age, the sex, the pre-treatment serum carcinoembryonic antigen level, the differentiation degree of the puncture tumor, the distance between the lower edge of the tumor and the anus for nuclear magnetic evaluation, the maximum diameter of the tumor, the grade of the invasion of blood vessels outside the wall, the invasion depth outside the tumor wall, the positive number of lymph nodes inside the mesentery, the positive number of lymph nodes on the side of the pelvic wall and the inguinal, the involvement condition of the fascia of the mesentery, whether the tumor is mucus adenocarcinoma and whether the patient receives the preoperative auxiliary treatment.
Further, the data preprocessing of the risk factor preprocessing module comprises a data cleaning unit, a data standardization unit and a single factor regression screening unit, wherein:
the data cleaning unit is used for cleaning the acquired data to obtain incomplete data, invalid data or error data and obtain data meeting the requirements;
the data standardization unit is used for processing the data obtained by the data cleaning unit and converting continuous data obtained by the data cleaning unit into classification or grade data according to a set standard;
the single-factor regression screening unit is used for extracting meaningful risk factors in single-factor analysis.
Furthermore, the single-factor regression screening unit extracts meaningful risk factors in single-factor analysis by analyzing the relationship between abnormal fluctuation of the single factor and tumor recurrence, metastasis or patient death within a certain time through a Kaplan-Meier product limit method in SPSS software.
Further, the multi-factor survival analysis is performed by a Cox regression reporting wizard model in SPSS software.
Further, the time period in the operator selection module for selection by the clinician ranges from 1 to 8 years.
Further, the content in the operator selection module for selection by the clinician includes a probability of tumor recurrence and a probability of patient death due to disease.
Furthermore, the survival model generation module further comprises an openness setting module, and the openness setting module is used for changing a program command according to the individualized requirement of a clinician and selecting a sample proportion of the randomly extracted and constructed model and an output probability scale value.
The invention also provides a prediction method of the survival prediction system for the T3-LARC patient before treatment, which comprises the following steps:
s1: the risk factor acquisition module acquires data of a T3-LARC patient, and transmits the data to the risk factor preprocessing module for preprocessing to obtain meaningful risk factors, specifically;
s1.1: the risk factor acquisition module acquires the age, the sex, the pre-treatment serum carcinoembryonic antigen level, the differentiation degree of a puncture tumor, the distance between the lower edge of the tumor and the anus for nuclear magnetic evaluation, the maximum diameter of the tumor, the invasion grade of blood vessels outside the wall, the invasion depth outside the wall of the tumor, the positive number of lymph nodes inside the mesentery, the positive number of lymph nodes on the side of the pelvic wall and the inguinal side, the involvement condition of mesentery fascia, whether the tumor is mucus adenocarcinoma and whether preoperative adjuvant therapy data are received;
s1.2: the risk factor acquisition module transmits the acquired information to the risk factor preprocessing module;
s1.3: the risk factor preprocessing module cleans the acquired information through a data cleaning unit to clean incomplete data, invalid data or error data to obtain data meeting the requirements;
s1.4: the data standardization unit of the risk factor preprocessing module processes the data obtained by the data cleaning unit, and the data standardization unit converts the continuous data obtained by the data cleaning unit into classification or grade data according to a set standard;
s1.5: a single-factor regression screening unit of the risk factor preprocessing module extracts meaningful risk factors in single-factor analysis by using data of the data standardization unit;
s2, the risk factor comprehensive analysis module further performs multi-factor survival analysis on the meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain the risk factors which can independently influence the survival of the patient;
s3: the survival model generation module generates and outputs a corresponding survival model, and specifically comprises:
s3.1: the survival model generation module receives information transmitted by the risk factor comprehensive analysis module and the operator selection module, wherein the operator selection module is used for a clinician to select contents to be predicted;
s3.2: the survival model generation module generates a corresponding survival model editing program through R-software according to the transmitted information and finally outputs a corresponding survival model;
s4: the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirements of the clinician.
In step S3, the survival model generation module further includes an openness setting module, and the openness setting module is configured to change the program command according to the personalized needs of the clinician, and select a sample proportion and an output probability scale value for randomly extracting the constructed model.
The invention has the beneficial effects that: the invention can provide individualized risk factor analysis of patients for clinical doctors before treatment, predict the recurrence rate and death rate of the patients in N years after operation, provide reference basis for the clinicians to formulate individualized treatment and follow-up schemes, have important clinical significance, and can greatly improve the prognosis of the patients, prolong the life cycle and improve the quality of life. Because some low-risk patients can be completely relieved clinically through the new auxiliary radiotherapy and chemotherapy before the operation, the observation follow-up mode of 'watch and wait' or the mode of partial excision under a mirror is avoided, and the patients who cannot protect the anus initially can obtain the chance of protecting the anus, so that the survival experience of the tumor patients is improved substantially. In addition, for a subset of patients with more risk factors and less predictive survival, clinical doctors are prompted to require a more systemic, more aggressive treatment regimen and a more intensive frequency of follow-up in order to extend the time interval between tumor recurrence and death and to take vigorous treatment in the first instance of recurrence.
At present, no model capable of combining image and clinical information prediction exists, and a part of models related to imaging omics cannot be popularized clinically at present.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a functional block diagram of a survival prediction system for a T3-LARC patient prior to treatment according to an embodiment of the present invention;
FIGS. 2A and 2B are each a nomogram of an embodiment of the present invention;
fig. 3A and 3B are verification set graphs corresponding to the 5-year recurrence prediction model and the 5-year death prediction model.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be derived by one of ordinary skill in the art from the embodiments given herein are intended to be within the scope of the present invention.
As shown in fig. 1-2, a survival prediction system for a T3-LARC patient before treatment according to an embodiment of the present invention includes a risk factor collecting module, a risk factor preprocessing module, a risk factor comprehensive analyzing module, an operator selecting module, a survival model generating module, and a survival rate displaying module, wherein,
the risk factor acquisition module is used for acquiring data of a T3-LARC patient;
the risk factor preprocessing module is used for preprocessing the data acquired by the risk factor acquisition module to obtain meaningful risk factors in single factor analysis;
the risk factor comprehensive analysis module is used for further performing multi-factor survival analysis on meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain risk factors capable of independently influencing the survival of the patient;
the operator selection module is used for the clinician to select the content required to be predicted;
the survival model generation module is used for receiving the information transmitted by the risk factor comprehensive analysis module and the operator selection module at the same time, generating a corresponding survival model editing program by using R-software, and finally outputting a corresponding survival model;
the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirement of an operator.
Further, the data collected by the risk factor collection module includes the patient's age, sex, pre-treatment serum carcinoembryonic antigen (CEA) level, degree of differentiation of the penetrating tumor, nuclear magnetic assessed distance between the tumor's lower margin and anus, maximum diameter of the tumor, grade of extramural vascular invasion (EMVI), depth of extramural invasion (EMD) and T3 subphase, number of intramesenteric lymph node positives, number of lateral and inguinal lymph nodes positives, rectal mesenteric fascia (MRF) involvement, whether the tumor is a mucinous adenocarcinoma, and whether it is subjected to pre-operative neoadjuvant therapy (NCRT).
Further, the data preprocessing of the risk factor preprocessing module comprises a data cleaning unit, a data standardization unit and a single factor regression screening unit, wherein:
the data cleaning unit is used for cleaning the acquired data to obtain incomplete data, invalid data or error data and obtain data meeting the requirements;
the data standardization unit is used for processing the data obtained by the data cleaning unit, and converting the continuous data obtained by the cleaning unit into classification or grade data according to a set standard;
the single-factor regression screening unit is used for extracting meaningful risk factors in single-factor analysis.
Furthermore, the single-factor regression screening unit extracts meaningful risk factors in single-factor analysis by analyzing the relationship between abnormal fluctuation of the single factor and tumor recurrence, metastasis or patient death within a certain time through a Kaplan-Meier product limit method in SPSS software.
Further, the multi-factor survival analysis is performed by a Cox regression reporting wizard model in SPSS software.
Further, the time period in the operator selection module for selection by the clinician ranges from 1 to 8 years.
Further, the content in the operator selection module for selection by the clinician includes a probability of tumor recurrence and a probability of patient death due to disease.
Furthermore, the survival model generation module further comprises an openness setting module, and the openness setting module is used for changing a program command according to the individualized requirement of an operator and selecting a sample proportion and an output probability scale value of the randomly extracted and constructed model.
The invention also provides a prediction method of the survival prediction system for the T3-LARC patient before treatment, which comprises the following steps:
s1: the risk factor acquisition module acquires data of a T3-LARC patient, and transmits the data to the risk factor preprocessing module for preprocessing to obtain meaningful risk factors, specifically;
s1.1: the risk factor acquisition module acquires the age, sex, pre-treatment serum carcinoembryonic antigen (CEA) level, differentiation degree of puncture tumor, nuclear magnetic evaluated distance between the lower edge of the tumor and anus, maximum diameter of the tumor, grade of extramural vascular invasion (EMVI), depth of extramural invasion (EMD) and T3 sub-stage, positive number of intramesenteric lymph nodes, positive number of lateral and inguinal lymph nodes of the pelvic wall, rectal mesenteric fascia (MRF) affected condition, whether the tumor is mucus adenocarcinoma and whether to receive preoperative neoadjuvant therapy (NCRT) of the T3-LARC patient;
s1.2: the risk factor acquisition module transmits the acquired information to the risk factor preprocessing module;
s1.3: the risk factor preprocessing module cleans the acquired information through a data cleaning unit to clean incomplete data, invalid data or error data to obtain data meeting the requirements;
s1.4: the data standardization unit of the risk factor preprocessing module processes the data obtained by the data cleaning unit, and the data standardization unit converts the continuous data obtained by the data cleaning unit into classification or grade data according to a set standard;
s1.5: a single-factor regression screening unit of the risk factor preprocessing module extracts meaningful risk factors in single-factor analysis by using data of the data standardization unit;
s2, the risk factor comprehensive analysis module further performs multi-factor survival analysis on the meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain the risk factors which can independently influence the survival of the patient;
s3: the survival model generation module generates and outputs a corresponding survival model, and specifically comprises:
s3.1: the survival model generation module receives information transmitted by the risk factor comprehensive analysis module and the operator selection module, wherein the operator selection module is used for a clinician to select contents to be predicted;
s3.2: the survival model generation module generates a corresponding survival model editing program through R-software according to the transmitted information and finally outputs a corresponding survival model;
s4: the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirements of the clinician.
In step S3, the survival model generation module further includes an openness setting module, and the openness setting module is configured to change the program command according to the personalized needs of the clinician, and select a sample proportion and an output probability scale value for randomly extracting the constructed model.
In order to facilitate understanding of the above-described technical aspects of the present invention, the above-described technical aspects of the present invention will be described in detail below in terms of specific usage.
The survival prediction system for the T3-LARC patient before treatment comprises a risk factor acquisition module, a risk factor preprocessing module, a risk factor comprehensive analysis module, an operator selection module, a survival model generation module and a survival rate display module, wherein the output end of the risk factor acquisition module is connected with the input end of the risk factor preprocessing module, the output end of the risk factor preprocessing module is connected with the input end of the risk factor comprehensive analysis module, the output end of the risk factor comprehensive analysis module and the output end of the operator selection module are connected with the input end of the survival model generation module, and the output end of the survival model generation module is connected with the input end of the survival rate display module.
The risk factor acquisition module is responsible for acquiring data of a T3-LARC patient, wherein the data comprise age, gender, pre-treatment serum carcinoembryonic antigen (CEA) level, differentiation degree of puncture tumor, nuclear magnetic evaluation distance between the lower edge of the tumor and anus, maximum diameter of the tumor, grade of extramural vascular invasion (EMVI), depth of extramural invasion (EMD) and T3 sub-stage, positive number of intramesenteric lymph nodes, positive number of lateral pelvic wall and inguinal lymph nodes, rectal mesenteric fascia (MRF) affected condition, whether the tumor is mucus adenocarcinoma and whether to receive preoperative new adjuvant therapy (NCRT).
The risk factor preprocessing module is responsible for preprocessing data and mainly comprises a data cleaning unit, a data standardization unit and a single-factor regression screening unit of the data, wherein the data cleaning unit is mainly responsible for cleaning the single-factor data to clean incomplete, invalid data and error data to obtain useful data, the data standardization unit is used for converting input continuous data into classified or graded data according to a set standard, different standards can be selected according to different requirements when the actual data standardization unit is applied, and the single-factor regression screening unit is mainly used for analyzing the relation between a single-factor continuous time variable and disease recurrence or transfer and death and extracting meaningful risk factors in single-factor analysis. The model mainly uses Kaplan-Meier product limit method in SPSS (Statistics 22; IBM Corp, Armonk, NY) software to analyze the relation between single-factor abnormal fluctuation and tumor recurrence or patient death within a certain time.
The risk factor comprehensive analysis module is responsible for carrying out multi-factor survival analysis on a meaningful risk factor, namely a factor P <0.05 (P is the probability of statistical hypothesis test) in the single-factor analysis obtained in the risk factor preprocessing module, and further carrying out multi-factor survival analysis through a Cox regression reporting hazard model in SPSS (Statistics 22; IBM Corp, Armonk, NY) software to obtain the risk factor (the factor P < 0.05) capable of obviously and independently influencing the survival of the patient.
The operator selection module is used for selecting the contents to be predicted mainly by a clinician, such as the probability of relapse within 3 years, the probability of death within 5 years and the like, wherein the selectable time range is 1-8 years, and the selectable prediction contents are the probability of relapse transfer or death.
The survival model generation module receives the information transmitted by the risk factor comprehensive analysis module and the operator selection module at the same time, generates a corresponding survival model editing program by using R-software (version 3.5.1; http:// www.Rproject.org), and finally outputs a corresponding survival model. The module also comprises an independent setting module which is used for changing the program command according to the requirement more individually by an operator and selecting the sample proportion of the random extraction building model and the output probability scale value.
The survival rate display module judges data according to the actual condition of the patient and makes corresponding output meeting the requirements of an operator.
The commands to which the present invention primarily relates are as follows:
# load rms packet
require(rms)
# import data
Csf = read.csv (file = "sample pointing instruction", header = T)
Setting the proportion of training samples, setting N% of samples as a training set, generating a model, using the rest 1-N% of samples as a verification set, and verifying the reliability of the samples
smp_size <- floor(N% * nrow(df))
Sampling randomly, X is random seed, calculating a series of random numbers according to the random seed, and sampling the random numbers as serial numbers
set.seed(X)
train_ind <- sample(seq_len(nrow(df)), size = smp_size)
# extraction training set and test set
train <- df[train_ind, ]
test <- df[-train_ind, ]
Calculate survival function from training set (one of the following is selected based on operator input, the former representing predicted recurrence and the latter representing predicted death)
S <- Surv(train$DFS,train$recurrence)
#S <- Surv(train$OS,train$death)
# setting the Factor to be analyzed (Factor is a Factor P <0.05 obtained from the Cox regression of the previous stage, auto-fill)
Factor0 = df$Factor0
Factor1 = df$Factor1
Factor2 = df$Factor2
……
Factor shown in # modified nomogram
ddist <- datadist(Factor0,Factor1,Factor2……)
options(datadist='ddist')
# multifactor regression
f <- cph(S ~ Factor0+Factor1+Factor2……,data=train,x=T,y=T,surv=T)
# acquisition time parameter
surv.cox <- Survival(f)
# time parameter for modification of prediction (N is input by the operator and represents the probability of recurrence or death occurring within N years of prediction)
surv1y <- function(x) (surv.cox(N,lp=x))
Display of # setup probability scale value (can be set by program operator through "setup" module according to required probability scale value)
ss <- c(0.01,0.05,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.85,0.9,0.95,0.99)
The # first parameter f is the cox regression results, the second parameter fun is the time set, funlabel is the label of the probability in the nomogram, and fun. Selecting the following two commands according to the selection of the operator, wherein N is the age selected by the operator, DFS is corresponding to the selection of the operator to predict recurrence, and OS is corresponding to the selection of the operator to predict death
mynom <- nomogram(f,fun = surv1y,funlabel = c("N year DFS probability"),fun.at = ss,lp = F)
#mynom <- nomogram(f,fun = surv1y,funlabel = c("N year OS probability"),fun.at = ss,lp = F)
Drawing nomogram
plot(mynom)
# training set survival function (one of the following is selected based on operator input, the former representing predicted relapse and the latter representing predicted death)
train_surv = Surv(train$DFS,train$recurrence)
#train_surv = Surv(train$OS,train$death)
Test set time parameter
test_surv=Surv(test$ DFS,test$ recurrence)
#test_surv=Surv(test$ OS,test$ death)
# calculating survivals of the training and test sets, times being the predicted probability of survival for a year, N being the predicted age entered by the operator
#estimates1=survest(f,newdata=train,times=N)$surv
estimates2=survest(f,newdata=test,times=N)$surv
# calculate consistency parameters for training and test sets
train_est <- rcorr.cens(x=estimates1,S=train_surv)
test_est <- rcorr.cens(x=estimates2,S=test_surv)
C Index and 95% CI of # training set
train_C <- train_est['C Index']
train_se <- train_est['S.D.']/2
train_hi <- train_C+1.96*train_se
train_low <- train_C-1.96*train_se
C Index and 95% CI of test set #
test_C <- test_est['C Index']
test_se <- test_est['S.D.']
test_hi <- test_C+1.96*test_se
test_low <- test_C-1.96*test_se
The present invention is described in detail below with reference to a specific example.
A sample containing 256T 3-LARC patients who visit more than 5 years at least from hospital A (the sample is attached in a folder by an excel. csv file) is used for constructing a Nomogram (https:// en. wikipedia. org/wiki/Nomogram) capable of predicting the 3-year recurrence rate and the fatality rate of the patients by applying the prediction system and the prediction method
https:// www.cnblogs.com/biostatisc/p/7903160. html).
Before data cleaning, sample data excel, csv is stored in a system D disc, and the system operates as follows:
1. the data cleaning unit is used for cleaning the acquired data to obtain incomplete data, invalid data or error data to obtain data meeting requirements, and the group of samples obtain excel1.csv after the data are cleaned
2. The data standardization unit is used for processing the data obtained by the data cleaning unit, and the processing comprises converting the continuous data obtained by the cleaning unit into classification or grade data according to a set standard to obtain excel2. csv;
3. risk factors which have significance on relapse and death in single-factor analysis are extracted through a Kaplan-Meier product limit method in SPSS software in a single-factor regression screening unit, wherein the risk factors include CEA level abnormality before treatment, biopsy tumor differentiation degree, tumor parts, maximum tumor diameter, extramural invasion depth, T3 sublevel, EMVI classification, suspicious lateral lymph nodes, MRF affected, mucus type rectal cancer and preoperative new adjuvant therapy (NCRT), and the factors without significant influence are removed to obtain excel3. csv.
4. And (3) carrying out multi-factor survival analysis on the meaningful factors in the step (3) by using a Cox regression developmental presentation wizard model in SPSS software of a risk factor comprehensive analysis unit, and further obtaining factors which have significance on relapse and death in the multi-factor survival analysis, wherein the factors include factors which have significance on relapse: pre-treatment serum CEA level abnormalities, EMVI staging, T3 subphase, MRF involvement, mucinous adenocarcinoma, pre-treatment NCRT, and factors that have significant effects on mortality: EMVI grading, T3 subphase, pre-treatment NCRT, and removal of factors without significant impact, yielded excel4. csv.
5. The prediction age is selected to be 5 years in the operator selection module, and the contents of the prediction are selected to be recurrence rate and disease death rate.
6. The probability scale value to be predicted is set through the openness setting module of the survival model generation module, and the probability scale value is set to be 0.99,0.95,0.9, 0.85,0.8, 0.7,0.6, 0.5,0.4,0.3,0.2,0.1, 0.05 and 0.01 at this time.
7. And clicking 'confirmation' to enter a survival model generation module, and automatically editing the internal language program by the R-software according to the file position and the input of an operator: (FIGS. 2A and 2B are graphs for recurrence prediction and death prediction, respectively)
A．
# load rms packet
require(rms)
# import data
df = read.csv(file = "excel4",header = T)
# sets the proportion of training samples, sets 70% of the samples as a training set, generates a model, sets the remaining 30% of the samples as a verification set, and verifies the reliability of the samples
smp_size <- floor(0.7 * nrow(df))
Randomly sampling, setting X as random seed, calculating a series of random numbers according to the random seed, and sampling with the random numbers as serial numbers
set.seed(X)
train_ind <- sample(seq_len(nrow(df)), size = smp_size)
# extraction training set and test set
train <- df[train_ind, ]
test <- df[-train_ind, ]
# calculation of survival function from training set (DFS disease-free survival, disease-free survival)
S <- Surv(train$DFS,train$recurrence)
# system assignment: setting the factors to be analyzed (factors with significant influence on relapse or fatality (P < 0.05) obtained from the previous stage of multifactorial survival analysis, Cox regression, auto-fill-in)
Factor0 = df $ mucinous adenocarcinoma
Factor1 = df $ MRF affected
Factor2 = df $ pre-treatment CEA abnormality
Factor3 = df $ T3 subperiod
Factor4 = df $ EMVI rating
Factor5 = df$NCRT
Factor shown in # modified nomogram
ddist <- datadist(Factor0,Factor1,Factor2,Factor3,Factor4,Factor5)
# multifactor regression
f <- cph(S ~ Factor0+Factor1+Factor2+Factor3+Factor4+Factor5,data=train,x=T,y=T,surv=T)
# acquisition time parameter
surv.cox <- Survival(f)
Time parameter for # modification prediction is "5" input by operator "
surv1y <- function(x) (surv.cox(5,lp=x))
Display of # set probability scale value
ss <- c(0.01,0.05,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.85,0.9,0.95,0.99)
The # first parameter f is the cox regression results, the second parameter fun is the time set, funlabel is the label of the probability in nomogram, and fun.
mynom <- nomogram(f,fun = surv1y,funlabel = c("5 year recurrence probability"),fun.at = ss,lp = F)
# rendering nomogram
plot(mynom)
# training set survival function
train_surv = Surv(train$DFS,train$recurrence)
Test set time parameter
test_surv=Surv(test$ DFS,test$ recurrence)
# calculating survivals of training and test sets, times being the predicted probability of survival for a year, 5 being the predicted age entered by the operator
estimates1=survest(f,newdata=train,times=5)$surv
estimates2=survest(f,newdata=test,times=5)$surv
# calculate consistency parameters for training and test sets
train_est <- rcorr.cens(x=estimates1,S=train_surv)
test_est <- rcorr.cens(x=estimates2,S=test_surv)
C Index and 95% CI of # training set
train_C <- train_est['C Index']
train_se <- train_est['S.D.']/2
train_hi <- train_C+1.96*train_se
train_low <- train_C-1.96*train_se
C Index and 95% CI of test set #
test_C <- test_est['C Index']
test_se <- test_est['S.D.']
test_hi <- test_C+1.96*test_se
test_low <- test_C-1.96*test_se
B.
# load rms packet
require(rms)
# import data
df = read.csv(file = "excel4",header = T)
# sets the proportion of training samples, sets 70% of the samples as a training set, generates a model, sets the remaining 30% of the samples as a verification set, and verifies the reliability of the samples
smp_size <- floor(0.7 * nrow(df))
Randomly sampling, setting X as random seed, calculating a series of random numbers according to the random seed, and sampling with the random numbers as serial numbers
set.seed(X)
train_ind <- sample(seq_len(nrow(df)), size = smp_size)
# extraction training set and test set
train <- df[train_ind, ]
test <- df[-train_ind, ]
# computation of survival function by training set (OS over survival, Total survival)
S <- Surv(train$OS,train$death)
# setting the factors to be analyzed (factors having a significant influence on relapse or mortality (P < 0.05) obtained from the previous stage of multifactorial survival analysis, Cox regression, auto-fill-in)
Factor3 = df $ T3 subperiod
Factor4 = df $ EMVI rating
Factor5 = df$NCRT
Factor shown in # modified nomogram
ddist <- datadist(Factor3,Factor4,Factor5)
options(datadist='ddist')
# multifactor regression
f <- cph(S ~ Factor3+Factor4+Factor5……,data=train,x=T,y=T,surv=T)
# acquisition time parameter
surv.cox <- Survival(f)
Time parameter for # modification prediction is "5" input by operator "
surv1y <- function(x) (surv.cox(5,lp=x))
Display of # set probability scale value
ss <- c(0.01,0.05,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.85,0.9,0.95,0.99)
The # first parameter f is the cox regression results, the second parameter fun is the time set, funlabel is the label of the probability in nomogram, and fun.
mynom <- nomogram(f,fun = surv1y,funlabel = c("5 year motality probability"),fun.at = ss,lp = F)
# rendering nomogram
plot(mynom)
# training set survival function
train_surv = Surv(train$OS,train$death)
Test set time parameter
test_surv=Surv(test$ OS,test$ death)
# calculating survivals of training and test sets, times being the predicted probability of survival for a year, 5 being the predicted age entered by the operator
estimates1=survest(f,newdata=train,times=5)$surv
estimates2=survest(f,newdata=test,times=5)$surv
# calculate consistency parameters for training and test sets
train_est <- rcorr.cens(x=estimates1,S=train_surv)
test_est <- rcorr.cens(x=estimates2,S=test_surv)
C Index and 95% CI of # training set
train_C <- train_est['C Index']
train_se <- train_est['S.D.']/2
train_hi <- train_C+1.96*train_se
train_low <- train_C-1.96*train_se
C Index and 95% CI of test set #
test_C <- test_est['C Index']
test_se <- test_est['S.D.']
test_hi <- test_C+1.96*test_se
test_low <- test_C-1.96*test_se
The corresponding nomograms output by the system are shown in fig. 2A and fig. 2B, with fig. 2A showing recurrence probability prediction and fig. 2B showing mortality prediction, and with Factor0-Factor5 representing the corresponding factors assigned in the system (see above). As 70% of samples are randomly drawn by the system to be used as prediction model generation samples, and the remaining 30% are used as prediction model verification samples, the system can simultaneously output C-index representing the reliability and repeatability of the model, and the index is greater than 0.8 to represent the clinically acceptable reliability. In this embodiment, the verification sets C-index corresponding to the 5-year recurrence prediction model and the 5-year death prediction model are 0.874 and 0.848, respectively, and as shown in fig. 3A and 3B, both are greater than 0.8, which indicates that the generated model can be reliably applied to the survival prediction of hospital a within a certain period (when the clinical treatment mode changes or the risk factors change, the system needs to be reused, and the treatment method and the corresponding risk factors need to be updated to perform model construction again). Using the above model, for patient B at visit, with elevated pre-treatment CEA serum levels, nuclear magnetic diagnosis of non-mucinous adenocarcinoma, MRF positive, EMVI grade 3, a T3 subperiod of 2, and need to receive NCRT, the total scores corresponding to the recurrence and mortality prediction nomograms were 130 and 115, respectively, corresponding to a 5-year recurrence rate of about 95% and a 5-year mortality rate of about 20% for that patient. For another patient C, normal pre-treatment CEA levels, nuclear magnetic diagnosis of mucinous adenocarcinoma, MRF negative, EMVI2 grade, T3 subperiod 1, and need to receive NCRT, the total score corresponding to recurrence and predictive of mortality is 70 and 45, respectively, corresponding to a 5-year recurrence rate of about 30% and a 5-year mortality rate of less than 5% in that patient. As described above, for each patient at a visit, a corresponding prediction can be made based on their risk factor condition. A score can be calculated according to the condition corresponding to each factor, and finally, a probability can be obtained by corresponding the obtained total score to the ruler table.
In conclusion, the invention can provide individualized risk factor analysis of patients for clinical doctors before treatment, predict the recurrence rate and death rate of the patients in N years after operation, provide reference basis for the clinicians to formulate individualized treatment and follow-up schemes, have important clinical significance, and can greatly improve the prognosis of the patients, prolong the life cycle and improve the quality of life. Because some low-risk patients can be completely relieved clinically through the new auxiliary radiotherapy and chemotherapy before the operation, the observation follow-up mode of 'watch and wait' or the mode of partial excision under a mirror is avoided, and the patients who cannot protect the anus initially can obtain the chance of protecting the anus, so that the survival experience of the tumor patients is improved substantially. In addition, for a subset of patients with more risk factors and less predictive survival, clinical doctors are prompted to require a more systemic, more aggressive treatment regimen and a more intensive frequency of follow-up in order to extend the time interval between tumor recurrence and death and to take vigorous treatment in the first instance of recurrence. The invention is suitable for being rapidly and widely developed in clinic and has extremely high application performance.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A survival prediction system aiming at a T3-LARC patient before treatment is characterized by comprising a risk factor acquisition module, a risk factor preprocessing module, a risk factor comprehensive analysis module, an operator selection module, a survival model generation module and a survival rate display module, wherein,
the risk factor acquisition module is used for acquiring data of a T3-LARC patient;
the risk factor preprocessing module is used for preprocessing the data acquired by the risk factor acquisition module to obtain meaningful risk factors in single factor analysis;
the risk factor comprehensive analysis module is used for further performing multi-factor survival analysis on meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain risk factors capable of independently influencing the survival of the patient;
the operator selection module is used for the clinician to select the content required to be predicted;
the survival model generation module is used for receiving the information transmitted by the risk factor comprehensive analysis module and the operator selection module at the same time, generating a corresponding survival model editing program by using R-software, and finally outputting a corresponding survival model;
the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirements of the clinician.
2. The system of claim 1, wherein the risk factor acquisition module acquires data including the patient's age, sex, pre-treatment serum carcinoembryonic antigen level, degree of differentiation of the punctured tumor, nuclear magnetic assessed distance between the lower margin of the tumor and the anus, maximum diameter of the tumor, grade of extramural vascular invasion, depth of extramural invasion of the tumor, number of positive intramesenteric lymph nodes, number of positive pelvic side and inguinal lymph nodes, involvement of the fascia of the rectal mesentery, whether the tumor is a mucinous adenocarcinoma, and whether pre-operative adjuvant therapy is required.
3. The survival prediction system of claim 1, wherein the risk factor pre-processing module pre-processes the data to comprise a data cleansing unit, a data normalization unit, and a one-way regression screening unit, wherein:
the data cleaning unit is used for cleaning the acquired data to obtain incomplete data, invalid data or error data and obtain data meeting the requirements;
the data standardization unit is used for processing the data obtained by the data cleaning unit and converting continuous data obtained by the data cleaning unit into classification or grade data according to a set standard;
the single-factor regression screening unit is used for extracting meaningful risk factors in single-factor analysis.
4. The survival prediction system of claim 3, wherein the single-factor regression screening unit is used for extracting significant risk factors from single-factor analysis by analyzing the relationship between abnormal fluctuation of single factors and tumor recurrence, metastasis or death of patients within a certain period of time through Kaplan-Meier product limit method in SPSS software.
5. The system of claim 1, wherein the multi-factor survival analysis is performed by a Cox regression clinical wizard model in the SPSS software.
6. The system of claim 1, wherein the time period selectable by the clinician in the operator selection module is in the range of 1-8 years.
7. The pre-treatment survival prediction system for a patient with T3-LARC according to claim 6, wherein the content selected by the clinician in the operator selection module comprises a probability of tumor recurrence and a probability of patient death due to illness.
8. The pre-treatment survival prediction system of claim 1, wherein the survival model generation module further comprises an openness setting module for modifying the program command according to the needs of the operator, and selecting the randomly extracted sample proportion of the constructed model and the output probability scale value.
9. A method of predicting survival of a patient with T3-LARC before treatment according to any one of claims 1 to 8, comprising the steps of:
s1: the risk factor acquisition module acquires data of a T3-LARC patient, and transmits the data to the risk factor preprocessing module for preprocessing to obtain meaningful risk factors, specifically;
s1.1: the risk factor acquisition module acquires the age, the sex, the pre-treatment serum carcinoembryonic antigen level, the differentiation degree of a puncture tumor, the distance between the lower edge of the tumor and the anus under nuclear magnetic evaluation, the maximum diameter of the tumor, the invasion grade of blood vessels outside the wall, the invasion depth outside the wall of the tumor, the positive number of lymph nodes inside the mesentery, the positive number of lymph nodes on the side of the pelvic wall and the inguinal side, the involvement condition of mesentery fascia, whether the tumor is mucus adenocarcinoma and whether preoperative adjuvant therapy is received;
s1.2: the risk factor acquisition module transmits the acquired information to the risk factor preprocessing module;
s1.3: the risk factor preprocessing module cleans the acquired information through a data cleaning unit to clean incomplete data, invalid data or error data to obtain data meeting the requirements;
s1.4: the data standardization unit of the risk factor preprocessing module processes the data obtained by the data cleaning unit, and the data standardization unit converts the continuous data obtained by the data cleaning unit into classification or grade data according to a set standard;
s1.5: a single-factor regression screening unit of the risk factor preprocessing module extracts meaningful risk factors in single-factor analysis by using data of the data standardization unit;
s2, the risk factor comprehensive analysis module further performs multi-factor survival analysis on the meaningful risk factors in the single-factor analysis obtained in the risk factor preprocessing module through software to obtain the risk factors which can independently influence the survival of the patient;
s3: the survival model generation module generates and outputs a corresponding survival model, and specifically comprises:
s3.1: the survival model generation module receives information transmitted by the risk factor comprehensive analysis module and the operator selection module, wherein the operator selection module is used for a clinician to select contents to be predicted;
s3.2: the survival model generation module generates a corresponding survival model editing program through R-software according to the transmitted information and finally outputs a corresponding survival model;
s4: the survival rate display module is used for judging the data of the survival model generation module according to the actual condition of the patient and making corresponding output meeting the requirements of the clinician.
10. The prediction method of survival prediction system for T3-LARC patient before treatment according to claim 9, wherein the survival model generation module further comprises an openness setting module in step S3, the openness setting module is configured to modify the program command according to the needs personalized by the clinician, and select the randomly extracted sample proportion of the constructed model and the output probability scale value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910017582.7A CN109599181B (en) | 2019-01-09 | 2019-01-09 | Survival prediction system and prediction method for T3-LARC patient before treatment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910017582.7A CN109599181B (en) | 2019-01-09 | 2019-01-09 | Survival prediction system and prediction method for T3-LARC patient before treatment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109599181A CN109599181A (en) | 2019-04-09 |
| CN109599181B true CN109599181B (en) | 2021-04-16 |
Family
ID=65965148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910017582.7A Active CN109599181B (en) | 2019-01-09 | 2019-01-09 | Survival prediction system and prediction method for T3-LARC patient before treatment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109599181B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110335679B (en) * | 2019-06-21 | 2020-05-05 | 山东大学 | Survival prediction method and system based on multi-granularity graph mode mining |
| CN110742633B (en) * | 2019-10-29 | 2023-04-18 | 慧影医疗科技(北京)股份有限公司 | Method and device for predicting risk after B-type aortic dissection operation and electronic equipment |
| CN111161882A (en) * | 2019-12-04 | 2020-05-15 | 深圳先进技术研究院 | Breast cancer life prediction method based on deep neural network |
| CN113160969A (en) * | 2021-04-14 | 2021-07-23 | 青岛大学附属医院 | Soft tissue sarcoma recurrence probability prediction method based on machine learning |
| CN114579674A (en) * | 2022-03-04 | 2022-06-03 | 山东汉图软件有限公司 | Method and system for judging user survival state based on user behavior |
| CN114664413B (en) | 2022-04-06 | 2022-12-20 | 中国医学科学院肿瘤医院 | System for predicting colorectal cancer treatment resistance and molecular mechanism thereof before treatment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101389957A (en) * | 2005-12-23 | 2009-03-18 | 环太平洋生物技术有限公司 | Prognosis prediction for colorectal cancer |
| CN106778046A (en) * | 2017-02-28 | 2017-05-31 | 思派（北京）网络科技有限公司 | The system and method that a kind of feature based value carries out prostate cancer recurrence prediction |
| CN107742534A (en) * | 2010-03-15 | 2018-02-27 | 新加坡保健服务集团有限公司 | Patient's viability forecasting system |
| WO2018148365A1 (en) * | 2017-02-09 | 2018-08-16 | Cognoa, Inc. | Platform and system for digital personalized medicine |
| CN108597603A (en) * | 2018-05-04 | 2018-09-28 | 吉林大学 | Cancer return forecasting system based on Multi-dimensional Gaussian distribution Bayes's classification |
| CN108922628A (en) * | 2018-04-23 | 2018-11-30 | 华北电力大学 | A kind of Prognosis in Breast Cancer survival rate prediction technique based on dynamic Cox model |
| CN109117864A (en) * | 2018-07-13 | 2019-01-01 | 华南理工大学 | Coronary heart disease risk prediction technique, model and system based on heterogeneous characteristic fusion |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180166174A1 (en) * | 2016-08-02 | 2018-06-14 | Thomas J. Lewis | Disease Mitigation and Elimination Health Learning Engine |
-
2019
- 2019-01-09 CN CN201910017582.7A patent/CN109599181B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101389957A (en) * | 2005-12-23 | 2009-03-18 | 环太平洋生物技术有限公司 | Prognosis prediction for colorectal cancer |
| CN107742534A (en) * | 2010-03-15 | 2018-02-27 | 新加坡保健服务集团有限公司 | Patient's viability forecasting system |
| WO2018148365A1 (en) * | 2017-02-09 | 2018-08-16 | Cognoa, Inc. | Platform and system for digital personalized medicine |
| CN106778046A (en) * | 2017-02-28 | 2017-05-31 | 思派（北京）网络科技有限公司 | The system and method that a kind of feature based value carries out prostate cancer recurrence prediction |
| CN108922628A (en) * | 2018-04-23 | 2018-11-30 | 华北电力大学 | A kind of Prognosis in Breast Cancer survival rate prediction technique based on dynamic Cox model |
| CN108597603A (en) * | 2018-05-04 | 2018-09-28 | 吉林大学 | Cancer return forecasting system based on Multi-dimensional Gaussian distribution Bayes's classification |
| CN109117864A (en) * | 2018-07-13 | 2019-01-01 | 华南理工大学 | Coronary heart disease risk prediction technique, model and system based on heterogeneous characteristic fusion |
Non-Patent Citations (3)
| Title |
|---|
| Can KRAS and BRAF mutations limit the benefit of liver resection in metastatic colorectal cancer patients;Passiglia F;《A systematic review and meta-analysis》;20161230;第99卷;第150-157页 * |
| 组学时代下机器学习方法在临床决策支持中的应用;赵学彤;《遗传》;20180730;第40卷(第9期);第693-703页 * |
| 结直肠癌肝转移患者术后生存的影响因素;胡翰杰;《肝癌电子杂志》;20160930;第3卷(第3期);第53-57页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109599181A (en) | 2019-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109599181B (en) | Survival prediction system and prediction method for T3-LARC patient before treatment | |
| Lange et al. | A joint model for multistate disease processes and random informative observation times, with applications to electronic medical records data | |
| KR20200077852A (en) | Medical image diagnosis assistance apparatus and method generating evaluation score about a plurality of medical image diagnosis algorithm | |
| JP2014516413A (en) | Cluster analysis of biomarker expression in cells | |
| JP2018068752A (en) | Machine learning device, machine learning method and program | |
| WO2020081582A1 (en) | Methods of diagnosing cancer using multiple artificial neural networks to analyze flow cytometry data | |
| CN105229471A (en) | For determining the system and method for preeclampsia risk based on biochemical biomarker analysis | |
| CN114242245A (en) | Machine learning method, system and device for predicting diabetic nephropathy occurrence risk based on electronic medical record data | |
| DE202016008673U1 (en) | Compositions for the active monitoring of prostate cancer | |
| CN115274118A (en) | Method for constructing testis tumor diagnosis and postoperative recurrence risk prediction model | |
| Camara et al. | Development of a convolutional neural network to detect abdominal aortic aneurysms | |
| Fan et al. | Cerebral microbleed automatic detection system based on the “deep learning” | |
| Low et al. | Automated detection of cecal intubation with variable bowel preparation using a deep convolutional neural network | |
| Selcuk et al. | Automated HER2 Scoring in Breast Cancer Images Using Deep Learning and Pyramid Sampling | |
| Gill et al. | Kidney Disease Classification Using Machine Learning Approach on DenseNet201 Model using Xray Images | |
| US9734122B2 (en) | System, method and computer-accessible medium for evaluating a malignancy status in at-risk populations and during patient treatment management | |
| Langham et al. | Predicting risk of dementia with machine learning and survival models using routine primary care records | |
| Kavya et al. | Heart Disease Prediction Using Logistic Regression | |
| JP2017516118A (en) | Noninvasive diagnostic method for early detection of fetal malformations | |
| CN114150059B (en) | MCM3 related breast cancer biomarker kit, diagnosis system and related application thereof | |
| Haque | Multiple myeloma. Detection from histological images using deep learning | |
| Saleena | Analysis of machine learning and deep learning prediction models for sepsis and neonatal sepsis: A systematic review | |
| Sangachin | A Statistical Decision Support System Incorporating Personalised Adaptive Reference Ranges for Longitudinal Monitoring in Prostate Cancer | |
| US20240311565A1 (en) | Methods and Systems for Identifying and Managing Biological Samples from Non-Human Subjects | |
| US20230031757A1 (en) | Health application |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |