CN109902421A - A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment - Google Patents
A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment Download PDFInfo
- Publication number
- CN109902421A CN109902421A CN201910176395.3A CN201910176395A CN109902421A CN 109902421 A CN109902421 A CN 109902421A CN 201910176395 A CN201910176395 A CN 201910176395A CN 109902421 A CN109902421 A CN 109902421A
- Authority
- CN
- China
- Prior art keywords
- cervical carcinoma
- carcinoma prognosis
- risk factor
- prognosis
- evaluation model
- 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
- 208000019065 cervical carcinoma Diseases 0.000 title claims abstract description 127
- 238000011156 evaluation Methods 0.000 title claims abstract description 24
- 238000004393 prognosis Methods 0.000 claims abstract description 120
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 48
- 230000004069 differentiation Effects 0.000 claims abstract description 37
- 238000000034 method Methods 0.000 claims abstract description 36
- 238000013210 evaluation model Methods 0.000 claims description 42
- 238000012549 training Methods 0.000 claims description 36
- 230000002792 vascular Effects 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 19
- 238000010276 construction Methods 0.000 claims description 13
- 238000010801 machine learning Methods 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000012706 support-vector machine Methods 0.000 claims description 8
- 208000035346 Margins of Excision Diseases 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 210000001165 lymph node Anatomy 0.000 claims description 5
- 230000005906 menstruation Effects 0.000 claims description 5
- 238000002271 resection Methods 0.000 claims description 5
- 108010088201 squamous cell carcinoma-related antigen Proteins 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 3
- 238000007473 univariate analysis Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 2
- 238000000151 deposition Methods 0.000 claims 1
- 208000014829 head and neck neoplasm Diseases 0.000 claims 1
- 208000026037 malignant tumor of neck Diseases 0.000 claims 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims 1
- 206010008342 Cervix carcinoma Diseases 0.000 abstract description 8
- 208000006105 Uterine Cervical Neoplasms Diseases 0.000 abstract description 8
- 201000010881 cervical cancer Diseases 0.000 abstract description 8
- 238000011282 treatment Methods 0.000 abstract description 6
- 230000002980 postoperative effect Effects 0.000 abstract description 3
- 230000004083 survival effect Effects 0.000 abstract description 3
- 238000002560 therapeutic procedure Methods 0.000 abstract description 2
- 238000009803 radical hysterectomy Methods 0.000 abstract 1
- 238000011127 radiochemotherapy Methods 0.000 abstract 1
- 230000008595 infiltration Effects 0.000 description 17
- 238000001764 infiltration Methods 0.000 description 17
- 206010041823 squamous cell carcinoma Diseases 0.000 description 12
- 201000010099 disease Diseases 0.000 description 8
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 8
- 238000002203 pretreatment Methods 0.000 description 6
- 201000011510 cancer Diseases 0.000 description 5
- 230000008859 change Effects 0.000 description 5
- 206010027476 Metastases Diseases 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011337 individualized treatment Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000001225 therapeutic effect Effects 0.000 description 2
- 210000001519 tissue Anatomy 0.000 description 2
- 201000009030 Carcinoma Diseases 0.000 description 1
- 208000005189 Embolism Diseases 0.000 description 1
- 208000007433 Lymphatic Metastasis Diseases 0.000 description 1
- 208000006374 Uterine Cervicitis Diseases 0.000 description 1
- 238000011353 adjuvant radiotherapy Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 206010008323 cervicitis Diseases 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000003211 malignant effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001575 pathological effect Effects 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 238000010837 poor prognosis Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The disclosure belongs to cervical carcinoma prognostic evaluation technical field, and in particular to a kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment.For I B- of FIGO, II A phase cervical carcinoma, Radical Hysterectomy and pelvic lymphadenectomy are a kind of effective treatment methods, postoperative five year survival rate is up to 87%~92%, but still having 10~20% patient that can recur, the patient for prognosis mala is the method for clinical common reply prognosis mala using postoperative chemoradiation therapy.For the lower problem of existing Sedlis evaluation criterion accuracy, present disclose provides a kind of models based on Sedlis standard, being predicted cervical cancer patient prognosis.For the model that the disclosure provides using tumor size, histology and differentiation degree as evaluation criterion, predictablity rate is 40.9%~92.0%, is better than Sedlis normative forecast accuracy.
Description
Technical field
The disclosure belongs to cervical carcinoma prognostic evaluation technical field, and in particular to a kind of cervical carcinoma based on Sedlis standard is pre-
The method for building up of model, cervical carcinoma prognostic system, storage medium and computer equipment afterwards.
Background technique
The information for disclosing the background technology part is merely intended to increase the understanding to the general background of the disclosure, without certainty
It is considered as recognizing or implying in any form that information composition has become existing skill well known to persons skilled in the art
Art.
Cervical carcinoma is the common malignant tumour of gynaecology, accounts for second in global female malignant.Now with uterine neck
The development of cancer census operations, the raising for early diagnosing technology, perfect and therapeutic scheme the rationalization of operation method, cervical carcinoma
Early diagnostic rate significantly improves earlier above, and the death rate constantly declines therewith.Although most cervical cancer patients can be cured through local treatment,
But still there is about 20% patient still because of recurrence and failure due to DISTANT METASTASES IN.Therefore, after postoperative adjuvant radiotherapy is widely used in operation
Patient with poor prognosis factor establishes the Preventive that the relevant prognostic model of cervicitis is conducive to calculate individual patients
Degree of danger is to instruct the clinical reasonable individualized treatment scheme of formulation, all to treatment income and survival rate for improving patient etc.
It is of great significance.
Currently, domestic and foreign scholars study more influence cervical cancer patient prognosis because being known as: the table of high-risk HPV, DNA
It reaches, age, clinical stages, lesion type, gross tumor volume (including tumor size and interstitial invasive depth), histological type, tissue
The molecular biology such as differentiation, infiltration by pathology palace, Carcinoma cell embolus, lymphatic metastasis, auxiliary therapy and SCC, CA125 refer to
Mark.In the factor that these influence prognosis, gross tumor volume, vascular and interstitial invasive depth are considered as the middle danger of cervical carcinoma prognosis
Factor.But single middle danger factor is not significant to the contributive rate of recurrence, middle danger factors in combination uses, and prediction recurrence rate can reach
15%.Later, it was tested according to prediction, GOG is combined into different middle danger factor groups according to factor of endangering in difference, proposes Sedlis
The infiltration of standard, i.e. vascular, interstitial infiltration deep 1/3, any tumor size;Vascular infiltration, interstitial infiltration in 1/3, tumor size >=
2cm;Vascular infiltration, interstitial infiltrate shallow 1/3, tumor size >=5cm;In no vascular infiltration, interstitial infiltration or deeply 1/3, tumour is big
Small >=4cm.Although Sedlis standard is not simply to be combined factor of endangering in 3, it is to the pre- of cervical cancer relapse
It is lower to survey accuracy, has the patients with recurrent close to half and does not meet Sedlis standard.
Since influence of the poor risk factor for cervical carcinoma prognosis is not fully clear, using different assessment moulds
Type and grading method bring result difference are larger, are unfavorable for holding disease wind in time for patient and medical staff
Danger, formulates effective therapeutic scheme.
Summary of the invention
For the studies above background, the disclosure is directed to the cervical cancer patient prognosis situation with moderate risk factor and provides
A kind of construction method of prognosis evaluation model.The disclosure for patient age, menstruation, final recognition, the phase not, histological type, point
Change degree, surgical resection margins, by palace, the influence factors such as vascular, interstitial, tumor size, lymph node and SCC-Ag screened, obtain
Risk factor relevant to cervical carcinoma prognosis, and precision of prediction analysis is carried out to the risk factor and cervical carcinoma Prognostic significance, it mentions
A kind of method for having supplied more accurate cervical carcinoma prognostic model, the model established using this method is with tumor size, histology
And differentiation degree is optimal combination, predictablity rate is 40.9%~92.0%, is better than Sedlis normative forecast accuracy.
In order to achieve the above technical purposes, the disclosure the following technical schemes are provided:
The disclosure is in a first aspect, provide a kind of construction method of cervical carcinoma prognosis evaluation model, the building of the assessment models
Method the following steps are included: establishing cervical carcinoma prognosis evaluation model, comment this using the method for machine learning by acquisition training sample
Estimate model to be trained;
The input factor of the cervical carcinoma prognosis evaluation model is risk factor levels, and output factor is cervical carcinoma prognosis feelings
Condition;
The training sample, comprising: the individual risk factor and corresponding cervical carcinoma prognosis situation of known cervical carcinoma individual;
The training: the risk factor of known cervical carcinoma individual is input in cervical carcinoma prognosis evaluation model, cervical carcinoma
Prognosis evaluation model exports cervical carcinoma prognosis prediction result;By the uterine neck of cervical carcinoma prognosis prediction result and known cervical carcinoma individual
Cancer prognosis situation carries out error calculation;If error is less than given threshold, training terminates, and obtains the cervical carcinoma prognosis that training finishes
Assessment models;If error is more than or equal to given threshold, training sample is updated again, continues to train, until error is less than setting
Threshold value.
Preferably, the machine learning method is SVM algorithm;It further, is the mode of Nonlinear Support Vector Machines.
Preferably, in the above method, the risk factor is selected from vascular (L), interstitial invasive depth (D), gross tumor volume
(T), histological type (H), three kinds or four kinds or five kinds of combination in differentiation degree (G).
Preferably, the risk factor further includes pre-treatment step, the pre-treatment step: being divided into vascular infiltration group and not
Infiltration group, vascular infiltration group are set as code name 1, and vascular does not infiltrate group and is set as 0;Interstitial invasive depth be divided into 1/3 group of invasive depth <,
1/3~2/3 group of invasive depth and 2/3 group of invasive depth >, 1/3 group of interstitial invasive depth < is set as 0, interstitial invasive depth 1/3
~2/3 group is set as 1,2/3 group of interstitial invasive depth > and is set as 2;The gross tumor volume gross tumor volume is divided into volume > 4cm group and body
Product≤4cm group, gross tumor volume > 4cm group are set as 1, and gross tumor volume≤4cm group is set as 0;The histological type is divided into squamous carcinoma group
With non-squamous carcinoma group, squamous carcinoma group is set as 0, and non-squamous carcinoma group is set as 1;The differentiation degree is divided into differentiated group, middle differentiation group and low point
Change group, differentiated are set as 2 being set as 0, middle differentiation group and be set as 1, low differentiation group.
It is further preferred that the risk factor is the group of gross tumor volume (T), histological type (H), differentiation degree (G)
It closes.
Disclosure second aspect, provides a kind of system for cervical carcinoma prognosis evaluation, which includes the root of computer
According to the device that test individual risk factor levels assess cervical carcinoma prognosis situation,
The computer includes database, memory, input unit, processor, display processing unit;
The input unit, for inputting the risk factor levels of test individual title and the individual;
The database is used to store the information of input unit input;
The memory is stored with the computer program that can be run on a processor;
The display processing unit, for exporting and showing prediction prognosis evaluation situation;
The processor calls the data in database, program in run memory, for by test individual it is dangerous because
Plain level is put into the cervical carcinoma prognosis situation that the test individual is obtained in the cervical carcinoma prognosis evaluation model after training;
The cervical carcinoma prognosis evaluation model that the training finishes is that a kind of input factor is risk factor levels, output factor
For the cervical carcinoma prognosis evaluation model of cervical carcinoma prognosis situation, which is established by the method for machine learning, using known sample
Individual risk factor and cervical carcinoma prognosis situation feature are trained in this.
Preferably, the machine learning method is SVM algorithm;It further, is the mode of Nonlinear Support Vector Machines.
Preferably, in the above method, the risk factor is selected from vascular (L), interstitial invasive depth (D), gross tumor volume
(T), histological type (H), three kinds or four kinds or five kinds of combination in differentiation degree (G).
Preferably, the risk factor further includes pre-treatment step, the pre-treatment step: being divided into vascular infiltration group and not
Infiltration group, vascular infiltration group are set as code name 1, and vascular does not infiltrate group and is set as 0;Interstitial invasive depth be divided into 1/3 group of invasive depth <,
1/3~2/3 group of invasive depth and 2/3 group of invasive depth >, 1/3 group of interstitial invasive depth < is set as 0, interstitial invasive depth 1/3
~2/3 group is set as 1,2/3 group of interstitial invasive depth > and is set as 2;The gross tumor volume gross tumor volume is divided into volume > 4cm group and body
Product≤4cm group, gross tumor volume > 4cm group are set as 1, and gross tumor volume≤4cm group is set as 0;The histological type is divided into squamous carcinoma group
With non-squamous carcinoma group, squamous carcinoma group is set as 0, and non-squamous carcinoma group is set as 1;The differentiation degree is divided into differentiated group, middle differentiation group and low point
Change group, differentiated are set as 2 being set as 0, middle differentiation group and be set as 1, low differentiation group.
It is further preferred that the risk factor is the group of gross tumor volume (T), histological type (H), differentiation degree (G)
It closes.
The disclosure third aspect provides the screening technique of risk factor described in first aspect or second aspect, this method packet
It includes following steps: preliminary screening being carried out by influence factor of the SPSS software to known individual and obtains m and cervical carcinoma prognosis phase
The risk factor of pass, wherein m > 3;
It is a combination from m-1, m risk factors of m risk factor sampling 3,4 ..., each combination does not repeat, and calculates every
The correlation of a combination and cervical carcinoma prognosis situation, selects the highest combination of correlation;
The known individual influence factor include patient age, menstruation situation, final recognition, phase not, histological type, differentiation
Degree, surgical resection margins, by palace, vascular, interstitial, tumor size, lymph node and SCC-Ag;
It is described that preliminary screening protocol is carried out as univariate analysis KM curve and multi-variables analysis Cox times by SPSS software
Return.
It is described calculate combination and the correlation of cervical carcinoma prognosis situation using Wilcoxo Signed-rank inspection come into
Row.
Disclosure fourth aspect provides a kind of storage medium for cervical carcinoma prognosis evaluation, has on the storage medium
The step of computer instruction, which is performed cervical carcinoma prognosis evaluation model building method as described in relation to the first aspect.
The 5th aspect of the disclosure, provides a kind of computer equipment for cervical carcinoma prognosis evaluation, the computer equipment packet
Include the computer instruction that memory, processor and storage execute on a memory and on a processor, the computer instruction
The step of cervical carcinoma prognosis evaluation model building method as described in relation to the first aspect is completed when running on a processor.
The disclosure the utility model has the advantages that
1. for the Sedlis technical problem lower as cervical carcinoma prognosis evaluation model accuracy used at present, this public affairs
It opens and provides the construction method of a kind of risk factor and cervical carcinoma prognosis status model, the model construction side provided using the disclosure
Method can obtain the higher cervical carcinoma prognostic evaluation model of precision of prediction, be conducive to obtain applied to clinical treatment more accurate
Prognostic evaluation provides reliable foundation as a result, formulating personalized therapy program for patient.
It is uterine neck 2. the disclosure additionally provides a kind of system predicted based on risk factor levels cervical carcinoma prognosis
Cancer prognosis evaluation provide can industrialized production product.
3. it is common at present more with the risk factor of evaluation cervical carcinoma prognosis situation, including patient age, menstruation, motherhood
Secondary, phase not, histological type, differentiation degree, surgical resection margins, by palace, vascular, interstitial, tumor size, lymph node and SCC-Ag
Deng.The disclosure additionally provides a kind of screening technique of risk factor, being capable of more exact evaluation cervical carcinoma prognosis feelings for screening
The risk factor of condition may filter out several and disease in influential factor to disease condition for numerous by SPSS first
The higher risk factor of situation degree of correlation, then the relationship of combination Yu the cervical carcinoma prognosis of each risk factor is investigated, to predict mould
The foundation of type and the foundation of prognostic system provide more accurate risk factor project
Detailed description of the invention
The Figure of description for constituting a part of this disclosure is used to provide further understanding of the disclosure, and the disclosure is shown
Meaning property embodiment and its explanation do not constitute the improper restriction to the disclosure for explaining the disclosure.
Fig. 1 is sample collection, training and model foundation flow chart in embodiment 2;
Fig. 2 is sample DFS and risk factor correlation results figure in embodiment 2
Fig. 3 is sample OS and risk factor correlation results figure in embodiment 2.
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the disclosure.Unless another
It indicates, all technical and scientific terms used herein has usual with disclosure person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the disclosure.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Leaving-one method (leave-one-out cross validation): leaving behind a sample does test set every time,
Its sample does training set, if there is k sample, then needs to train k times, test k times.
Disease-free survival (DFS): first operation to first time recurrence (clinical or radiological evidence or no evidence of disease patient
Last time medical day).
Total existence (OS): the day of first operation stops to dead or follow-up in last day.
As background technique is introduced, although most cervical cancer patients can be cured through local treatment, but still have about 20%
Patient still because of recurrence and failure due to DISTANT METASTASES IN.Therefore, the prognosis situation of cervical cancer patient is assessed, is taken accordingly
Individualized treatment scheme helps to improve treatment income.Clinical hazard factor assessment model Sedlis prognosis used at present is commented
It is lower to estimate accuracy, the disclosure is directed to the technical problem, studies the relationship of patient risk's factor level and prognosis, provides
The construction method and cervical carcinoma prognostic system of a kind of cervical carcinoma prognosis evaluation model is established with gross tumor volume, tissue
It learns type, the model that differentiation degree is evaluation criterion, effectively increases the accuracy to the prediction of patient's prognosis situation.
In order to enable those skilled in the art can clearly understand the technical solution of the disclosure, below with reference to tool
The technical solution of the disclosure is described in detail in the embodiment and comparative example of body.
Embodiment 1
In the present embodiment, provide the specific implementation case screened to risk factor: inventor has collected Shandong doctor
The clinical and pathological data of nearly IB-IIA more than 10 year, cervical cancer patient phase of institute acquires the influence factor information of known individual, the shadow
The factor of sound include: patient age, menstruation, final recognition, phase not, histological type, differentiation degree, surgical resection margins, by palace, vascular,
Interstitial, tumor size, lymph node and SCC-Ag etc.), and using SPSS software by variable analysis it is preliminary go out it is pre- with cervical carcinoma
The step of possible related factor afterwards, screening is 19. version of SPSS using prognosis (dead or recurrence difference) as endpoints,
Univariate analysis KM curve, it is believed that the variable of P < 0.05 is significant risk factor, and then multi-variables analysis Cox is returned, P
The factor of < 0.05 is the independent hazard factor about prognosis.
Embodiment 2
By preliminary screening, obtain five kinds of risk factors relevant to cervical carcinoma prognosis situation, respectively vascular (L),
Interstitial invasive depth (D), gross tumor volume (T), histological type (H), differentiation degree (G).It is pre- in order to establish effective cervical carcinoma
Assessment models afterwards, inventor are predicted using machine learning algorithm.Training set according to above-mentioned risk factor is a considerable number of, becomes
Fewer feature is measured, sample is trained and is predicted by the way of Nonlinear Support Vector Machines.It endangers first from above-mentioned five kinds
Three kinds, four kinds and five kinds factors are chosen in dangerous factor respectively to combine as one, by the group of the risk factor levels of known individual
Conjunction is fitted with the individual cervical carcinoma prognostic level, establishes that input factor is risk factor levels, output factor is cervical carcinoma
The cervical carcinoma prognosis evaluation model of prognosis situation has the statistics of the supporting vector machine model of accuracy by appended sample assessment
Conspicuousness.Inventor examines the forecasting accuracy come between the different model groups of comparison using Wilcoxo Signed-rank.Pass through
The statistical result of computation model, using gross tumor volume (T), histological type (H), differentiation degree (G) combination as it is dangerous because
Element and the prognosis situation of cervical carcinoma have better accuracy.
According to the risk factor of selection, external certificate is carried out to the prognosis of cervical carcinoma using SVM algorithm.To survive, final result is
Example, inventor is first by a large amount of case application in training and study.Since the dead event number of patient (n=44) is less than total event
Number, inventor randomly selects without dead event patient 44, to avoid the bias of two class patients.Then 88 patients are randomly divided into
Training group and experimental group.Model wherein is established by training group of 87 patients, another 1 patient is external certificate model prediction training
The accuracy of group.100 different combinations are repeated in this process, such as 12 kinds of combination S edlis standard replacements, every time to accurate
Property is calculated.By finding the statistical significance of model, more accurate prediction model is inventors herein proposed.
After the completion of model training, further pass through the extension of cross validation, one of them (or multiple) sample in the present embodiment
Originally it is excluded except entire model development, feature selecting and models fitting, except models fitting.When not only
When vertical test set is available, which can effectively make up the insufficient defect of sample size.Leave- is used in the present embodiment
Selected model is verified in the verifying of one-out cross-module type.This process repeats 100 times, and it is accurate to calculate final result prediction every time
Degree.
The accuracy rate of Sedlis normative forecast DFS is between 36.4% to 71.6%.With other 12 kinds of combined forecasting precisions
It compares, has 8 kinds of combined forecasting precisions higher, difference is statistically significant.Tumor size, histology and differentiation degree are best group
It closes, predictablity rate is 40.9%~92.0%.The repetition predictablity rate of Sedlis standard is between 37.5% to 60.0%.
10 kinds of combined prediction results are superior to Sedlis standard, have statistical significance.The highest accuracy rate of prediction model is that tumour is big
The synthesis of small, histology and differentiation degree, accuracy rate are 42.6%~86.8%.
Embodiment 3
A kind of construction method of cervical carcinoma prognosis evaluation model, the construction method packet of the assessment models are provided in the present embodiment
It includes following steps: establishing cervical carcinoma prognosis evaluation model, training sample is obtained, using the method for machine learning to the assessment models
It is trained;
The input factor of the cervical carcinoma prognosis evaluation model is risk factor levels, and output factor is cervical carcinoma prognosis feelings
Condition;
The training sample, comprising: the individual risk factor and corresponding cervical carcinoma prognosis situation of known cervical carcinoma individual;
The training: the risk factor of known cervical carcinoma individual is input in cervical carcinoma prognosis evaluation model, cervical carcinoma
Prognosis evaluation model exports cervical carcinoma prognosis prediction result;By the uterine neck of cervical carcinoma prognosis prediction result and known cervical carcinoma individual
Cancer prognosis situation carries out error calculation;If error is less than given threshold, training terminates, and obtains the cervical carcinoma prognosis that training finishes
Assessment models;If error is more than or equal to given threshold, training sample is updated again, continues to train, until error is less than setting
Threshold value.
In the present embodiment, the machine learning method is the mode of Nonlinear Support Vector Machines.
In the present embodiment, the combination of the risk factor gross tumor volume (T), histological type (H), differentiation degree (G).
In the present embodiment, the pre-treatment step: the risk factor further includes pre-treatment step, the pretreatment step
It is rapid: to be divided into vascular infiltration group and do not infiltrate group, vascular infiltration group is set as code name 1, and vascular does not infiltrate group and is set as 0;Interstitial infiltration is deep
Degree is divided into 1/3 group of invasive depth <, 1/3~2/3 group of invasive depth and 2/3 group of invasive depth >, 1/3 group of interstitial invasive depth <
It is set as 1/3~2/3 group of 0, interstitial invasive depth and is set as 1,2/3 group of interstitial invasive depth > being set as 2;The gross tumor volume tumour body
Integral is volume > 4cm group and volume≤4cm group, and gross tumor volume > 4cm group is set as 1, and gross tumor volume≤4cm group is set as 0;It is described
Histological type is divided into squamous carcinoma group and non-squamous carcinoma group, and squamous carcinoma group is set as 0, and non-squamous carcinoma group is set as 1;The differentiation degree is divided into high score
Change group, middle differentiation group and low differentiation group, differentiated are set as 2 being set as 0, middle differentiation group and be set as 1, low differentiation group.
Embodiment 4
In the present embodiment, one kind is provided, test individual cervical carcinoma prognosis situation is carried out based on appraisal procedure in embodiment 2
The system of prediction, the system include being assessed according to test individual risk factor levels cervical carcinoma prognosis situation for computer
Device,
The computer includes database, memory, input unit, processor, display processing unit;
The input unit, for inputting the risk factor levels of test individual title and the individual;
The database is used to store the information of input unit input;
The memory is stored with the computer program that can be run on a processor;
The display processing unit, for exporting and showing prediction prognosis evaluation situation;
The processor calls the data in database, program in run memory, for by test individual it is dangerous because
Plain level is put into the cervical carcinoma prognosis situation that the test individual is obtained in the cervical carcinoma prognosis evaluation model after training;
The cervical carcinoma prognosis evaluation model that the training finishes is that a kind of input factor is risk factor levels, output factor
For the cervical carcinoma prognosis evaluation model of cervical carcinoma prognosis situation, which is established in a manner of Nonlinear Support Vector Machines, is used
Individual risk factor and cervical carcinoma prognosis situation feature are trained in known sample.
In the present embodiment, the risk factor is gross tumor volume, histological type, differentiation degree, acquisition device acquisition disease
The gross tumor volume information of people, histological type, differentiation degree are translated into computer-readable language.
Embodiment 5
In the present embodiment, a kind of storage medium for cervical carcinoma prognosis evaluation is provided, has on the storage medium and calculates
The step of machine instruction, which is done as follows appraisal procedure:
It is obtained for the risk factor levels of test individual to be put into the cervical carcinoma prognosis evaluation model after training
The cervical carcinoma prognosis situation of the test individual;
The cervical carcinoma prognosis evaluation model that the training finishes is that a kind of input factor is risk factor levels, output factor
For the cervical carcinoma prognosis evaluation model of cervical carcinoma prognosis situation, which is established in a manner of Nonlinear Support Vector Machines, is used
Individual risk factor and cervical carcinoma prognosis situation feature are trained in known sample.
In the present embodiment, the risk factor is gross tumor volume, histological type, differentiation degree, acquisition device acquisition disease
The gross tumor volume information of people, histological type, differentiation degree are translated into computer-readable language.
Embodiment 6
In the present embodiment, a kind of computer equipment for cervical carcinoma prognosis evaluation is provided, which includes
The computer instruction that memory, processor and storage execute on a memory and on a processor, the computer instruction exist
The step of completing following appraisal procedure when running on processor:
It is obtained for the risk factor levels of test individual to be put into the cervical carcinoma prognosis evaluation model after training
The cervical carcinoma prognosis situation of the test individual;
The cervical carcinoma prognosis evaluation model that the training finishes is that a kind of input factor is risk factor levels, output factor
For the cervical carcinoma prognosis evaluation model of cervical carcinoma prognosis situation, which is established in a manner of Nonlinear Support Vector Machines, is used
Individual risk factor and cervical carcinoma prognosis situation feature are trained in known sample.
In the present embodiment, the risk factor is gross tumor volume, histological type, differentiation degree, acquisition device acquisition disease
The gross tumor volume information of people, histological type, differentiation degree are translated into computer-readable language.
The foregoing is merely preferred embodiment of the present disclosure, are not limited to the disclosure, for the skill of this field
For art personnel, the disclosure can have various modifications and variations.It is all within the spirit and principle of the disclosure, it is made any to repair
Change, equivalent replacement, improvement etc., should be included within the protection scope of the disclosure.
Claims (10)
1. a kind of construction method of cervical carcinoma prognosis evaluation model, which is characterized in that the construction method of assessment models includes following
Step: establishing cervical carcinoma prognosis evaluation model, obtains training sample, is instructed using the method for machine learning to the assessment models
Practice;
The input factor of the cervical carcinoma prognosis evaluation model is risk factor levels, and output factor is cervical carcinoma prognosis situation;
The training sample, comprising: the individual risk factor and corresponding cervical carcinoma prognosis situation of known cervical carcinoma individual;
The training: the risk factor of known cervical carcinoma individual is input in cervical carcinoma prognosis evaluation model, cervical carcinoma prognosis
Assessment models export cervical carcinoma prognosis prediction result;Cervical carcinoma prognosis prediction result and the cervical carcinoma of known cervical carcinoma individual is pre-
Situation carries out error calculation afterwards;If error is less than given threshold, training terminates, and obtains the cervical carcinoma prognosis evaluation that training finishes
Model;If error is more than or equal to given threshold, training sample is updated again, continues to train, until error is less than given threshold.
2. the construction method of cervical carcinoma prognosis evaluation model as described in claim 1, which is characterized in that the machine learning side
Method is SVM algorithm.
3. the construction method of cervical carcinoma prognosis evaluation model as described in claim 1, which is characterized in that the machine learning side
Method is the mode of Nonlinear Support Vector Machines.
4. the construction method of cervical carcinoma prognosis evaluation model as described in claim 1, which is characterized in that the risk factor is
Three kinds or four kinds or five kinds of the combination in vascular, interstitial invasive depth, gross tumor volume, histological type, differentiation degree.
5. the construction method of cervical carcinoma prognosis evaluation model as claimed in claim 4, which is characterized in that the risk factor is
The combination of gross tumor volume, histological type, differentiation degree.
6. a kind of system for cervical carcinoma prognosis evaluation, which is characterized in that the system comprises computers according to be measured
The device that body risk factor levels assess cervical carcinoma prognosis situation,
The computer includes database, memory, input unit, processor, display processing unit;
The input unit, for inputting the risk factor levels of test individual title and the individual;
The database is used to store the information of input unit input;
The memory is stored with the computer program that can be run on a processor;
The display processing unit, for exporting and showing prediction prognosis evaluation situation;
The processor calls the data in database, program in run memory, for by the risk factor water of test individual
Lay flat the cervical carcinoma prognosis situation that the test individual is obtained in the cervical carcinoma prognosis evaluation model after training;
The cervical carcinoma prognosis evaluation model that the training finishes be a kind of input factor be risk factor levels, output factor is palace
The cervical carcinoma prognosis evaluation model of neck cancer prognosis situation, which is established by the method for machine learning, using in known sample
Individual risk factor and cervical carcinoma prognosis situation feature are trained.
7. the screening technique of risk factor described in any one of claim 1-5 or claim 6, which is characterized in that described
Method obtains m and cervical carcinoma the following steps are included: carrying out preliminary screening by influence factor of the SPSS software to known individual
The relevant risk factor of prognosis, wherein m > 3;
It is a combination from m-1, m risk factors of m risk factor sampling 3,4 ..., each combination does not repeat, and calculates each group
The correlation with cervical carcinoma prognosis situation is closed, the highest combination of correlation is selected;
The known individual influence factor includes patient age, menstruation situation, final recognition, the phase is other, histological type, breaks up journey
Degree, surgical resection margins, by palace, vascular, interstitial, tumor size, lymph node and SCC-Ag;
It is described that preliminary screening protocol is carried out as univariate analysis KM curve and multi-variables analysis Cox recurrence by SPSS software.
8. the method for claim 7, which is characterized in that the calculating is combined and the correlation of cervical carcinoma prognosis situation is
Progress is examined by Wilcoxo Signed-rank.
9. a kind of storage medium for cervical carcinoma prognosis evaluation, which is characterized in that refer on the storage medium with computer
It enables, which is performed the step of any one of the processing claim 1-5 cervical carcinoma prognosis evaluation model building method
Suddenly.
10. a kind of computer equipment for for cervical carcinoma prognosis evaluation, which is characterized in that the computer equipment includes depositing
The computer instruction that reservoir, processor and storage execute on a memory and on a processor, the computer instruction are being located
The step of any one of the processing claim 1-5 cervical carcinoma prognosis evaluation model building method is performed when running on reason device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910176395.3A CN109902421A (en) | 2019-03-08 | 2019-03-08 | A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910176395.3A CN109902421A (en) | 2019-03-08 | 2019-03-08 | A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109902421A true CN109902421A (en) | 2019-06-18 |
Family
ID=66946645
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910176395.3A Pending CN109902421A (en) | 2019-03-08 | 2019-03-08 | A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109902421A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110634571A (en) * | 2019-09-20 | 2019-12-31 | 四川省人民医院 | Prognosis prediction system after liver transplantation |
CN111640518A (en) * | 2020-06-02 | 2020-09-08 | 山东大学齐鲁医院 | Cervical cancer postoperative survival prediction method, system, equipment and medium |
CN111640509A (en) * | 2020-06-02 | 2020-09-08 | 山东大学齐鲁医院 | Cervical cancer postoperative recurrence risk prediction method and system |
CN112233123A (en) * | 2020-09-21 | 2021-01-15 | 广州金域医学检验中心有限公司 | Method and device for classifying tumor stroma in pathological picture, computer equipment and storage medium |
WO2021143781A1 (en) * | 2020-01-14 | 2021-07-22 | 之江实验室 | Multi-center synergetic cancer prognosis prediction system based on multi-source migration learning |
CN113736882A (en) * | 2021-09-08 | 2021-12-03 | 河北医科大学第二医院 | Gene marker for screening early cervical cancer and product and application thereof |
CN113838570A (en) * | 2021-08-31 | 2021-12-24 | 华中科技大学 | Cervical cancer self-consistent typing method and system based on deep learning |
CN114496306A (en) * | 2022-01-28 | 2022-05-13 | 北京大学口腔医学院 | Machine learning-based prognosis survival stage prediction method and system |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130337449A1 (en) * | 2010-12-13 | 2013-12-19 | Samsung Life Public Welfare Foundation | Marker for predicting gastric cancer prognosis and method for predicting gastric cancer prognosis using the same |
CN108389624A (en) * | 2018-01-31 | 2018-08-10 | 浙江大学 | Duodenal adenocarcinoma prognosis prediction system based on COX proportional hazard models |
-
2019
- 2019-03-08 CN CN201910176395.3A patent/CN109902421A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130337449A1 (en) * | 2010-12-13 | 2013-12-19 | Samsung Life Public Welfare Foundation | Marker for predicting gastric cancer prognosis and method for predicting gastric cancer prognosis using the same |
CN108389624A (en) * | 2018-01-31 | 2018-08-10 | 浙江大学 | Duodenal adenocarcinoma prognosis prediction system based on COX proportional hazard models |
Non-Patent Citations (2)
Title |
---|
杜成凤等: "基于支持向量机模型预测结直肠癌根治术预后风险", 《重庆医科大学学报》 * |
林晓桃等: "宫颈癌患者预后因素分析及预测模型的初步建立", 《中国医药导报》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110634571A (en) * | 2019-09-20 | 2019-12-31 | 四川省人民医院 | Prognosis prediction system after liver transplantation |
WO2021143781A1 (en) * | 2020-01-14 | 2021-07-22 | 之江实验室 | Multi-center synergetic cancer prognosis prediction system based on multi-source migration learning |
US11456078B2 (en) | 2020-01-14 | 2022-09-27 | Zhejiang Lab | Multi-center synergetic cancer prognosis prediction system based on multi-source migration learning |
CN111640518A (en) * | 2020-06-02 | 2020-09-08 | 山东大学齐鲁医院 | Cervical cancer postoperative survival prediction method, system, equipment and medium |
CN111640509A (en) * | 2020-06-02 | 2020-09-08 | 山东大学齐鲁医院 | Cervical cancer postoperative recurrence risk prediction method and system |
CN112233123A (en) * | 2020-09-21 | 2021-01-15 | 广州金域医学检验中心有限公司 | Method and device for classifying tumor stroma in pathological picture, computer equipment and storage medium |
CN112233123B (en) * | 2020-09-21 | 2024-05-14 | 广州金域医学检验中心有限公司 | Tumor interstitial classification method, device, computer equipment and storage medium in pathological pictures |
CN113838570A (en) * | 2021-08-31 | 2021-12-24 | 华中科技大学 | Cervical cancer self-consistent typing method and system based on deep learning |
CN113838570B (en) * | 2021-08-31 | 2024-04-26 | 华中科技大学 | Cervical cancer self-consistent typing method and system based on deep learning |
CN113736882A (en) * | 2021-09-08 | 2021-12-03 | 河北医科大学第二医院 | Gene marker for screening early cervical cancer and product and application thereof |
CN114496306A (en) * | 2022-01-28 | 2022-05-13 | 北京大学口腔医学院 | Machine learning-based prognosis survival stage prediction method and system |
CN114496306B (en) * | 2022-01-28 | 2022-12-20 | 北京大学口腔医学院 | Machine learning-based prognosis survival stage prediction method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109902421A (en) | A kind of cervical carcinoma prognostic evaluation methods, system, storage medium and computer equipment | |
Shariat et al. | Nomograms for bladder cancer | |
Tomatis et al. | Late rectal bleeding after 3D-CRT for prostate cancer: development of a neural-network-based predictive model | |
Tantipalakorn et al. | Outcome and patterns of recurrence for International Federation of Gynecology and Obstetrics (FIGO) stages I and II squamous cell vulvar cancer | |
CN108694718A (en) | The same period new chemoradiation therapy curative effect evaluation system and method before rectal cancer | |
CN109124660B (en) | Gastrointestinal stromal tumor postoperative risk detection method and system based on deep learning | |
CN107561280B (en) | A kind of kit for predicting breast cancer relapse | |
CN109055562A (en) | A kind of biomarker, predict clear-cell carcinoma recurrence and mortality risk method | |
CN105574356A (en) | Breast tumor modeling and diagnostic method | |
CN109599181A (en) | A kind of Prediction of survival system and prediction technique being directed to T3-LARC patient before the treatment | |
CN111027548A (en) | Brain glioma molecular marker nondestructive prediction method and system based on image omics | |
Mathur et al. | Detecting malignancy of ovarian tumour using convolutional neural network: A review | |
Zhu et al. | Individualized prediction of metastatic involvement of lymph nodes posterior to the right recurrent laryngeal nerve in papillary thyroid carcinoma | |
CN108363900A (en) | Based on the pyrosequencing result judgement patients with gliomas of MGMT promoter methylations to the system of Temozolomide sensibility | |
Uccella et al. | Survival in clinical stage I endometrial cancer with single vs. multiple positive pelvic nodes: results of a multi-institutional Italian study | |
KR20200109544A (en) | Multi-cancer classification method by common significant genes | |
RU2563437C1 (en) | Method for prediction of clinical outcome of breast cancer | |
CN105243300A (en) | Approximation spectral clustering algorithm based method for predicting cancer metastasis and recurrence | |
RU2523138C1 (en) | Method for prediction of risk of developing progression of disease following radiofrequency thermoablation of hepatic metastases from colorectal cancer | |
Stone | The dawning of the age of artificial intelligence in urology | |
Liu et al. | Nomograms for predicting long‐term overall survival and cancer‐specific survival in chordoma: a population‐based study | |
Pati et al. | Multiplexed tumor profiling with generative AI accelerates histopathology workflows and improves clinical predictions | |
Kang et al. | A Nomogram for Predicting Three or More Axillary Lymph Node Involvement before Breast Cancer Surgery | |
Lagano et al. | Novel morphological tools in predicting pancreatic neuroendocrine tumors (Pan-Net) clinical outcome | |
Taylor et al. | Dynamic optical contrast imaging as a novel modality to rapidly distinguish oral squamous cell carcinoma from surrounding normal tissue |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190618 |