CN115862850B - Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data - Google Patents
Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data Download PDFInfo
- Publication number
- CN115862850B CN115862850B CN202310154793.1A CN202310154793A CN115862850B CN 115862850 B CN115862850 B CN 115862850B CN 202310154793 A CN202310154793 A CN 202310154793A CN 115862850 B CN115862850 B CN 115862850B
- Authority
- CN
- China
- Prior art keywords
- model
- follow
- sample
- information
- longitudinal
- 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
Images
Abstract
The invention discloses a modeling method and a device of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, and relates to the technical field of medical treatment, wherein the method comprises the following steps: acquiring longitudinal follow-up case information of a sample patient, and extracting sample evaluation information from the longitudinal follow-up case information; determining first sample characteristic information from the sample evaluation information, and constructing a first model based on the first sample characteristic information; inputting sample evaluation information of a sample patient into a first model to obtain a first prediction score output by the first model; determining second sample characteristic information from the first prediction score and the sample evaluation information of the sample patient, and constructing a second model based on the second sample characteristic information and the first prediction score; and sequentially processing the first model and the second model to construct a third model. The third model constructed based on longitudinal multidimensional data can better monitor the hepatocellular carcinoma occurrence condition of a patient, and has a good effect when being applied to liver cirrhosis crowds.
Description
Technical Field
The invention relates to the technical field of medical treatment, in particular to a modeling method and device of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data.
Background
Hepatocellular carcinoma is a major histological subtype of liver cancer, and can account for over 90% of primary liver cancer.
If the monitoring of the hepatocellular carcinoma can be more accurately carried out, the patients with high risk of the hepatocellular carcinoma can be identified early, the method is helpful for guiding clinicians to select and implement more accurate treatment measures for the patients, and the patients can be subjected to intervention and treatment as early as possible, so that better operation effects are obtained, the prognosis of the patients is improved, and the method is one of key measures capable of remarkably improving the survival rate of the hepatocellular carcinoma.
Currently, there are a variety of hepatocellular carcinoma predictive models available for hepatocellular carcinoma monitoring in clinic, however, the efficacy of these models is often significantly reduced in the population with cirrhosis. Because only single follow-up data is used, most of the models are baseline clinical data, most of the models lack real data for monitoring the efficacy of the hepatocellular carcinoma in a long term, the extremely early occurrence condition and the progress change of the hepatocellular carcinoma cannot be correctly reflected, and the clinical problems of unreasonable monitoring and management of the hepatocellular carcinoma, untimely treatment and management, and untimely communication between doctors and patients and the like are easily caused.
Disclosure of Invention
In view of the above, the embodiment of the invention provides a modeling method and a device for a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, so as to solve the problem that the hepatocellular carcinoma prediction model cannot accurately monitor hepatocellular carcinoma.
According to a first aspect, an embodiment of the present invention provides a method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, the method comprising:
longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information;
determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
determining second sample feature information from the first predictive score and the sample assessment information for the sample patient, and constructing a second model based on the second sample feature information and the first predictive score for the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
Sequentially processing the constructed first model and the constructed second model to construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
With reference to the first aspect, in a first implementation manner of the first aspect, the acquiring longitudinal follow-up case information of the sample patient, processing the longitudinal follow-up case information, and extracting sample evaluation information of the sample patient from the longitudinal follow-up case information specifically includes the following steps:
acquiring the longitudinal follow-up case information of the sample patient, and determining the type of the longitudinal follow-up case information;
if the longitudinal follow-up case information is determined to be unstructured type information, carrying out structuring treatment on the unstructured longitudinal follow-up case information to obtain structured longitudinal follow-up case information;
screening from the structured longitudinal follow-up case information, and extracting the sample evaluation information.
With reference to the first aspect, in a second implementation manner of the first aspect, the first model is obtained by modeling through the following steps:
Determining an initial evaluation factor corresponding to the first sample characteristic information;
adjusting the initial evaluation factor to obtain a first evaluation factor;
the first model is constructed based on the first sample characteristic information and the first evaluation factor.
With reference to the second implementation manner of the first aspect, in a third implementation manner of the first aspect, the inputting the sample evaluation information of the sample patient into the constructed first model, to obtain a first prediction score corresponding to the sample patient output by the first model specifically includes:
inputting the sample evaluation information of a sample patient into the first model, and processing the sample evaluation information based on the first sample characteristic information and the first evaluation factor corresponding to the first model to obtain the first prediction score output by the first model.
With reference to the second implementation manner of the first aspect, in a fourth implementation manner of the first aspect, the second model is obtained by modeling through the following steps:
grouping the sample patients to obtain a sample patient of a diseased group and a sample patient of a non-diseased group;
fitting the sample evaluation information of the sample patients in the diseased group and the non-diseased group and a first prediction score generated by a previous follow-up based on an initial longitudinal model to respectively obtain a first group average contour of the diseased group and a second group average contour of the non-diseased group; the initial longitudinal model adopts a multi-element linear mixed effect model;
The second model is constructed based on the first set of average contours and the second set of average contours.
With reference to the fourth implementation manner of the first aspect, in a fifth implementation manner of the first aspect, when the third model is applied, the second prediction score is obtained by:
when a first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds a first score, longitudinal follow-up visit case information of the person to be evaluated in the next follow-up visit and the previous follow-up visit and the first predictive score of each follow-up visit are obtained, the longitudinal follow-up visit case information is processed, and evaluation information of the person to be evaluated is extracted from the longitudinal follow-up visit case information;
determining second characteristic information of the current follow-up visit from the first prediction scores corresponding to the multiple follow-up visits of the person to be evaluated and the evaluation information;
inputting the first predictive scores and the evaluation information of the multiple follow-up visits into the second model, and calculating a longitudinal profile corresponding to the person to be evaluated based on the initial longitudinal model;
comparing the longitudinal profile with the first group average profile and the second group average profile respectively to obtain a first grouping probability of the evaluators to be evaluated into the diseased group and a second grouping probability of the evaluators to be evaluated into the non-diseased group;
And calculating the second prediction score corresponding to the to-be-evaluated person based on the first grouping probability.
With reference to the first aspect, in a sixth implementation manner of the first aspect, when the third model is applied, the third prediction score is obtained by:
longitudinal follow-up case information of a person to be evaluated is obtained, the longitudinal follow-up case information is processed, and evaluation information of the person to be evaluated is extracted from the longitudinal follow-up case information;
determining first characteristic information of the first follow-up visit from the evaluation information of the first follow-up visit;
inputting the first characteristic information of the first follow-up visit into the first model to obtain a first prediction score corresponding to the first follow-up visit of the person to be evaluated output by the first model;
determining second characteristic information of the current follow-up visit from the first predictive score of the current follow-up visit and the evaluation information of the previous follow-up visit of the person to be evaluated when the first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds a first score;
inputting the second characteristic information of the current follow-up visit into the second model to obtain the second prediction score corresponding to the current follow-up visit of the evaluator output by the second model, and obtaining a third prediction score of the evaluator based on the first prediction score of the first follow-up visit and the second prediction score of the current follow-up visit;
And when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person is not more than the first score, obtaining a third predictive score of the to-be-evaluated person based on the first predictive score of the first follow-up visit.
In a second aspect, an embodiment of the present invention further provides a modeling apparatus for a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, the apparatus comprising:
the information extraction module is used for acquiring longitudinal follow-up case information of the sample patient, processing the longitudinal follow-up case information and extracting sample evaluation information of the sample patient from the longitudinal follow-up case information;
a first construction module for determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
the prediction evaluation module is used for inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
A second construction module for determining second sample characteristic information from the first predictive score and the sample evaluation information of the sample patient, and constructing a second model based on the second sample characteristic information and the first predictive score of the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
a third construction module, configured to sequentially process the constructed first model and the constructed second model, and construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data as described in any one of the above when the program is executed.
In a fourth aspect, embodiments of the present invention also provide a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of a method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data as described in any one of the above.
According to the modeling method and device for the hepatocellular carcinoma monitoring model based on the longitudinal multidimensional data, the sample evaluation information of which the longitudinal data type comprises the serological markers is extracted from the longitudinal follow-up case information of a sample patient, the first model is built based on the sample evaluation information, the second model is built based on the sample evaluation information and the first prediction scores output by the first model, and finally the final third model is obtained based on the first model and the second model through sequential application, so that the method and device have higher hepatocellular carcinoma monitoring precision and good layering monitoring effect of the hepatocellular carcinoma under the condition of not increasing extra economic burden, doctors can be guided to make better treatment schemes for the patient through the third prediction scores output by the third model, and particularly, the third model still has good effect when being applied to a population with cirrhosis, and feasibility is increased for practical popularization and application in clinic.
Drawings
The features and advantages of the present invention will be more clearly understood by reference to the accompanying drawings, which are illustrative and should not be construed as limiting the invention in any way, in which:
fig. 1 shows a flow diagram of a modeling method of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data provided by the invention;
Fig. 2 is a schematic flow chart of acquiring sample evaluation information in the modeling method of the hepatocellular carcinoma monitoring model based on longitudinal multidimensional data;
fig. 3 is a schematic flow chart of a second prediction score obtaining process in the modeling method of the hepatocellular carcinoma monitoring model based on longitudinal multidimensional data;
fig. 4 is a schematic flow chart of a third prediction score obtaining process in the modeling method of the hepatocellular carcinoma monitoring model based on longitudinal multidimensional data provided by the invention;
fig. 5 shows a schematic structural diagram of a modeling method device of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data;
fig. 6 shows a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
Hepatocellular carcinoma is a major histological subtype of liver cancer, and can account for over 90% of primary liver cancer. After the chronic liver disease patient finishes treatment in the hospital, the hospital can acquire the longitudinal follow-up case information with time sequence characteristics of the patient at different time points in a follow-up mode, and timely obtain relevant feedback of the patient condition change. Some patients may have multiple diseases, and the follow-up mode can also help doctors to track and observe information data of other diseases of the patients to a certain extent. Good prognosis of hepatocellular carcinoma depends in large part on early diagnosis and disease stage of hepatocellular carcinoma. The average survival time of early/very early liver cancer patients after surgical excision, ablation surgery and liver transplantation is more than 5 years.
If the monitoring of the hepatocellular carcinoma can be more accurately carried out, the patients with high risk of the hepatocellular carcinoma can be identified early, the method is helpful for guiding clinicians to select and implement more accurate treatment measures for the patients, and the patients can be subjected to intervention and treatment as early as possible, so that better operation effects are obtained, the prognosis of the patients is improved, and the method is one of key measures capable of remarkably improving the survival rate of the hepatocellular carcinoma.
Currently, there are a variety of hepatocellular carcinoma predictive models for hepatocellular carcinoma monitoring clinically, but due to the lack of multi-etiology, multi-clinical feature construction and verification queues, and the current complex clinical trends of more and more chronic liver disease patients with more complications, metabolic disorders, antiviral treatment effects, the application of the hepatocellular carcinoma predictive model is limited and the efficacy of the hepatocellular carcinoma predictive model in the population needs to be further evaluated. Because only single follow-up data is based on baseline clinical data, the traditional COX or Logistic regression model and other hepatocellular carcinoma prediction models ignore the characteristic that the hepatocellular carcinoma risk/recurrence risk is progressed/resolved in the follow-up time, namely the characteristic of continuous change, and the efficacy of the hepatocellular carcinoma prediction model is obviously reduced in the population of patients suffering from chronic liver diseases such as liver cirrhosis. Clinically, the problems of unreasonable monitoring and management of the hepatocellular carcinoma, untimely treatment and management, and communication between doctors and patients are also easy to cause.
The follow-up visit refers to an observation method for a hospital to regularly know the change of the illness state of a patient and guide the rehabilitation of the patient in a communication or other modes for the patient who has been in the visit, the service level before and after the hospital can be improved through the follow-up visit, meanwhile, a doctor can conveniently track and observe the patient, and the first hand data is mastered to carry out statistical analysis and accumulated experience, so that the patient can be better served.
In order to solve the above-mentioned problems, as shown in fig. 1, fig. 1 is a flow chart of a modeling method of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data provided in the present embodiment. The modeling method of the embodiment of the invention can be used in electronic equipment. As shown in fig. 1, the modeling method includes the steps of:
s10, longitudinal follow-up case information of the sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information. The machine direction may characterize the chronological order, i.e. the time sequence. The longitudinal follow-up case information characterizes demographic characteristics, clinical characteristics, laboratory examination indicators, and medical and medication history of the sample patient at various time points of the follow-up process with time series characteristics.
Longitudinal follow-up case information for a sample patient may be obtained by way of follow-up. The longitudinal follow-up case information can be stored in a database of a medical and health system of a hospital, and can also be text information and image information which are fed back by a sample patient independently, and can also be obtained after medical staff records the text information and the image information which are fed back by the sample patient, for example, the relevant information recorded on a case book. The method for acquiring the longitudinal follow-up case information is not limited at all, and only the longitudinal follow-up case information can be acquired.
Since the longitudinal follow-up case information may be unstructured data, in step S10, the unstructured data in the longitudinal follow-up case information is also structured, so that sample evaluation information of the evaluator can be extracted therefrom. It should be noted that not all the structured longitudinal follow-up case information is sample evaluation information, and the sample evaluation information is related data/features for subsequent model modeling, which are obtained after screening from the structured longitudinal follow-up case information.
In the embodiment of the invention, the sample evaluation information of the sample patient is clinical data of the sample patient, including gender, age, liver index information and the like, and the liver index information can be obtained by a medical physical examination mode and the like, and comprises the following information:
Alpha Fetoprotein (AFP), albumin (ALB), total Bilirubin (TBIL), and platelet count (PLT) in sample patients.
S20, determining first sample characteristic information from sample evaluation information of a sample patient, and constructing a first model based on the first sample characteristic information of the sample patient, wherein the first model is used for predicting a first prediction score, and the first prediction score is used for representing a hepatocellular carcinoma monitoring condition corresponding to the sample patient.
In the embodiment of the invention, the first prediction score is obtained based on a first model, wherein the first model adopts an aMAP score model, and the first model is an aMAP (age-Male-ALBI-Platelets) model.
In practical application, the input data of the first model is evaluation information extracted from current follow-up case information generated by follow-up of the to-be-evaluated person, and the evaluation information is used for obtaining a first prediction score of the to-be-evaluated person.
The first sample characteristic information includes part of information in sex, age, and liver index information of the sample patient. More specifically, the aamap model computes a corresponding first predictive score based on the gender, age, albumin, total bilirubin, and platelet count of the sample patient in liver index information, i.e., the first sample characteristic information includes age, gender, albumin, total bilirubin, and platelet count.
That is, the first model contains only three common clinical test indexes, namely Albumin (ALB), total Bilirubin (TBIL) and platelet count (PLT), besides age and gender, so that the risk prediction model and the method based on the risk prediction model can be well popularized in primary hospitals, and the aap score can obtain better evaluation performance in patients in liver inflammation activity period because the index of liver inflammation is not involved.
In an embodiment of the invention, the first model is modeled by:
a10, determining the first sample characteristic information.
The determining process of the first sample characteristic information can be seen in detail in step S20 shown in fig. 1, and will not be described in detail herein.
a20, determining an initial evaluation factor corresponding to the first sample characteristic information.
a30, adjusting the initial evaluation factor to obtain a first evaluation factor.
a40, constructing a first model based on the first sample characteristic information and the first evaluation factor.
The calculation mode of the finally obtained first prediction score is as follows:
aMAP= ({ 0.06×age+0.89×gender (male: 1; female: 0) +0.48× [ (log) 10 TBIL× 0.66) + (ALB × -0.085)] -0.01 ×PLT)} + 7.4)/14.77 × 100
It is understood that 0.06, 0.89, 0.48, 0.66, -0.085, 0.01, 7.4, 14.77 and 100 in the above formula are all the corresponding first evaluation factors of the aap scoring model.
S30, inputting sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model.
Specifically, step S30 includes: sample evaluation information of the sample patient is input into a first model, and the sample evaluation information is processed based on first sample characteristic information and a first evaluation factor corresponding to the first model to obtain a first prediction score output by the first model.
S40, determining second sample characteristic information from the first prediction scores and the sample evaluation information of the sample patients, and constructing a second model based on the second sample characteristic information and the first prediction scores of the sample patients, wherein the second model is used for predicting second prediction scores, and the second prediction scores are also used for representing the hepatocellular carcinoma monitoring conditions corresponding to the sample patients.
In the embodiment of the invention, the second prediction score is obtained based on a second model, and the second model also utilizes an aMAP score model, and the second model is an aMAP-2 model.
In practical application, the input data of the second model is evaluation information and a first prediction score extracted from longitudinal follow-up case information generated by the follow-up of the person to be evaluated at the time of the follow-up and the previous follow-up, and the second model is used for obtaining a second prediction score of the person to be evaluated.
The second sample characteristic information includes part of information in liver index information of the sample patient. More specifically, the aap-2 model calculates a corresponding second predictive score based on the first predictive score of the sample patient, alpha fetoprotein in the liver index information, i.e., the second sample characteristic information includes the first predictive score and alpha fetoprotein.
In an embodiment of the invention, the second model is modeled by the following steps:
b10, grouping the sample patients to obtain a sample patient of a diseased group (hepatocellular carcinoma group patient) and a sample patient of a non-diseased group (non-hepatocellular carcinoma group patient).
b20, fitting sample evaluation information of sample patients in the diseased group and the non-diseased group and a first prediction score generated during the follow-up and the previous follow-up based on an initial longitudinal model to obtain a first group average contour of the diseased group and a second group average contour of the non-diseased group respectively, wherein the initial longitudinal model adopts a multi-element linear mixed effect model (Multivariate Linear Mixed Effect Models, MLMM).
In the embodiment of the invention, the structure of the initial longitudinal model is as follows:
wherein, when the MLMM is fitted,
，/>
representing a patient, ++>
Representing a certain follow-up of the patient, XAndZdifferent design matrices representing fitting variables, E representing desired values,/->
And->
Representing the fixed coefficients and random coefficients in the linear mixture effect model of the fitting variables respectively,MVNis a multi-element normal distribution. />
Measurement and calculation using Bayesian Markov chain Monte Carlo algorithm (Markov-ChainMonte-Carlo, MCMC), +.>
For calculating the variance and the model internal parameters in covariance.
b30, constructing a second model based on the first group of average contours and the second group of average contours.
Non-and hepatocellular carcinoma group patients are denoted by numerals 0 and 1, respectively, parameters
The production, wherein,
sample patients who have participated in the second model modeling enter a training library, and each time a new sample patient and its sample evaluation information are acquired, the longitudinal profile of the new sample patient is calculated by using the sample evaluation information, and compared with the first group average profile and the second group average profile to give a group prediction probability that the sample patient is classified as a hepatocellular carcinoma group, namely, a second prediction score (aMAP-2 score), which is used for continuing training the second model. The range of values for the second predictive score is 0-1, the closer the second predictive score is to 1, the greater the probability that the sample patient is classified as a group of hepatocellular carcinoma.
In the embodiment of the invention, the grouping prediction probability is obtained based on a marginal prediction method, namely, the aMAP-2 score is obtained based on the marginal prediction method, and a specific calculation formula is as follows:
in the formula (i),
predicting probability for a packet->
Representing marginal prediction method, < >>
Representing a patient,/->
Is group, i.e. diseased group or non-diseased group, < ->
For the addition sign, the ∈10>
，/>
Is a vector of length 2, +.>
A priori probability weight ratio representing that the sample patient belongs to each group,/->
Representing parameters generated during model learning.
For each bayesian MCMC simulation, two chains of sufficient length were used (6000 iterations total, the first 3000 iterations burned out) and the convergence between the two MCMC chains was evaluated by the trajectory graph, all bayesian MCMC simulations using no information priors generated by the algorithm.
Dynamic changes in hepatocellular carcinoma risk during patient follow-up are demonstrated by the changed aap-2 score that will be generated for each follow-up, and the average time difference between the time that a hepatocellular carcinoma patient is predicted to be a high risk outcome by an aap-2 score and its clinical diagnosis can be represented by the average prediction period (MLT). MLT results of aMAP-2 scores show that hepatocellular carcinoma patients can be early warned for the first time to be at high risk of hepatocellular carcinoma, so that monitoring of the hepatocellular carcinoma is enhanced in time and intervention is performed in time.
S50, sequentially processing the constructed first model and the constructed second model to construct a third model. The third model is used for predicting a third prediction score, and the third prediction score is also used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
In the embodiment of the invention, the third prediction score is obtained based on a third model, and the third model is an aamap aMAP-2 model because the third model is obtained based on sequential application of the constructed first model and the second model. Specifically, during actual application of the third model, sequential application of the first model and the second model means that a first prediction score of the to-be-evaluated person is obtained by using the first model, and when the first prediction score corresponding to the first follow-up visit of the to-be-evaluated person exceeds the first score, a second prediction score of the to-be-evaluated person is obtained by using the second model, and a third prediction score of the to-be-evaluated person is obtained according to the previous first prediction score and the second prediction score; and when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person does not exceed the first score, obtaining a third predictive score directly according to the first predictive score.
Therefore, the third predictive score obtained based on the third model has higher hepatocellular carcinoma monitoring precision and good layering monitoring effect on hepatocellular carcinoma, and can guide doctors to develop better treatment schemes for patients.
According to the modeling method of the hepatocellular carcinoma monitoring model, the sample evaluation information of which the longitudinal data type comprises the serological markers is extracted from the longitudinal follow-up case information of the sample patient, the first model is constructed based on the sample evaluation information, the second model is constructed based on the sample evaluation information and the first prediction score output by the first model, and finally the final third model is obtained based on the first model and the second model by sequential application, so that the method has higher hepatocellular carcinoma monitoring precision and good layering monitoring effect of the hepatocellular carcinoma under the condition of not increasing extra economic burden, and the third prediction score output by the third model can guide doctors to formulate better treatment schemes for patients.
In order to solve the above-mentioned problems, as shown in fig. 2, in this embodiment, a method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data is provided, where the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention may be used in an electronic device, fig. 2 is a schematic flow chart of obtaining sample evaluation information of the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention, as shown in fig. 2, step S10 includes the following steps:
S11, longitudinal follow-up case information of the sample patient is acquired, and the type of the longitudinal follow-up case information is determined.
S12, determining the longitudinal follow-up case information as unstructured type information, and carrying out structuring treatment on the unstructured longitudinal follow-up case information to obtain structured longitudinal follow-up case information.
S13, screening from the structured longitudinal follow-up case information, and extracting sample evaluation information.
Unstructured data in the longitudinal follow-up case information is structured in step S12, so that sample evaluation information of an evaluator can be extracted therefrom, and sample evaluation information for subsequent model modeling, which is obtained after screening from the structured longitudinal follow-up case information, is obtained in step S13.
In order to solve the above-mentioned problems, as shown in fig. 3, a method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data is provided in this embodiment, and the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention may be used in an electronic device, and fig. 3 is a schematic flow chart of a second prediction score obtaining process in the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention, as shown in fig. 3, where the second prediction score is obtained by:
A10, when the first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds the first score, acquiring the longitudinal follow-up visit case information of the person to be evaluated in the current follow-up visit and the previous follow-up visit, processing the longitudinal follow-up visit case information, and extracting the evaluation information of the person to be evaluated from the longitudinal follow-up visit case information.
The manner of acquiring the evaluation information in step a10 is detailed in step S10 shown in fig. 1, and is not described herein.
A20, determining second characteristic information from the first prediction scores and the evaluation information corresponding to the multiple follow-up visits of the person to be evaluated.
The determining manner of the second feature information is detailed in step S40 shown in fig. 1, and is not described herein.
A30, inputting the first predictive scores and the evaluation information of the multiple follow-up visits into the second model, and calculating the longitudinal profile corresponding to the person to be evaluated based on the initial longitudinal model.
Step a30 is detailed in the modeling process of the second model, and will not be described herein.
And A40, comparing the longitudinal profile with the first group average profile and the second group average profile respectively to obtain a first grouping probability of the evaluators to be generalized as a diseased group and a second grouping probability of the evaluators to be generalized as a non-diseased group.
Similarly, step a40 is to obtain a first packet probability and a second packet probability based on a marginal prediction method.
A50, calculating a second prediction score corresponding to the person to be evaluated based on the first grouping probability.
In order to solve the above-mentioned problems, as shown in fig. 4, in this embodiment, a method and an apparatus for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data are provided, where the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention may be used in an electronic device, and fig. 4 is a schematic flow chart of a third prediction score obtaining process in the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention, as shown in fig. 4, the third prediction score is obtained by:
and B10, acquiring longitudinal follow-up case information of the to-be-evaluated person, processing the longitudinal follow-up case information, and extracting evaluation information of the to-be-evaluated person from the longitudinal follow-up case information.
Step B10 is detailed in step S10 shown in FIG. 1, and is not described herein.
B20, determining first characteristic information of the first follow-up from the evaluation information of the first follow-up.
And B30, inputting the first characteristic information of the first follow-up visit into the first model to obtain a first prediction score corresponding to the first follow-up visit of the person to be evaluated, which is output by the first model.
Step B30 is detailed in step S30 shown in FIG. 1, and is not described herein.
And B40, determining second characteristic information of the current follow-up visit from the first predictive score of the current follow-up visit and the previous follow-up visit of the to-be-evaluated person and evaluation information when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person exceeds the first score.
The determination manners of the first feature information and the second feature information are respectively detailed in step S20 and step S40 shown in fig. 1, and are not described herein.
And B50, inputting the second characteristic information of the current follow-up visit into a second model to obtain a second prediction score corresponding to the current follow-up visit of the to-be-evaluated person output by the second model, and obtaining a third prediction score of the to-be-evaluated person based on the first prediction score of the first follow-up visit and the second prediction score of the current follow-up visit.
And B60, when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person is determined not to exceed the first score, obtaining a third predictive score of the to-be-evaluated person based on the first predictive score corresponding to the first follow-up visit.
The modeling method device of the hepatocellular carcinoma monitoring model based on the longitudinal multi-dimensional data, which is described below, and the modeling method of the hepatocellular carcinoma monitoring model based on the longitudinal multi-dimensional data, which is described above, can be correspondingly referred to each other.
As shown in fig. 5, in this embodiment, a method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data is provided, where the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention may be used in an electronic device, and fig. 5 is a schematic structural diagram of the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data according to the embodiment of the present invention, as shown in fig. 5, where the apparatus includes:
the information extraction module 10 is configured to obtain longitudinal follow-up case information of a sample patient, process the longitudinal follow-up case information, and extract sample evaluation information of the sample patient from the longitudinal follow-up case information. The longitudinal follow-up case information is information recording the demographics, clinical characteristics, laboratory examination indexes, disease history, drug administration history and the like of the sample patient.
Longitudinal follow-up case information for a sample patient may be obtained by way of follow-up. The longitudinal follow-up case information can be stored in a database of a medical and health system of a hospital, and can also be text information and image information which are fed back by a sample patient independently, and can also be obtained after medical staff records the text information and the image information which are fed back by the sample patient, for example, the relevant information recorded on a case book. The method for acquiring the longitudinal follow-up case information is not limited at all, and only the longitudinal follow-up case information can be acquired.
Since the longitudinal follow-up case information may be unstructured data, the unstructured data in the longitudinal follow-up case information is also structured in the first building module 20, so that sample evaluation information of the evaluator can be extracted therefrom. It should be noted that not all the structured longitudinal follow-up case information is sample evaluation information, and the sample evaluation information is related data/features for subsequent model modeling, which are obtained after screening from the structured longitudinal follow-up case information.
In the embodiment of the invention, the sample evaluation information of the sample patient is clinical data of the sample patient, including gender, age, liver index information and the like, and the liver index information can be obtained by medical physical examination and the like, and the liver index information comprises Alpha Fetoprotein (AFP), albumin (ALB), total Bilirubin (TBIL) and platelet count (PLT) of the sample patient.
The first construction module 20 is configured to determine first sample characteristic information from sample evaluation information of a sample patient, and construct a first model based on the first sample characteristic information of the sample patient, where the first model is used for predicting a first prediction score, and the first prediction score is used for characterizing a hepatocellular carcinoma monitoring condition corresponding to the sample patient.
In the embodiment of the invention, the first prediction score is obtained based on a first model, the first model adopts an aMAP score model, and the first model is an aMAP model.
The first sample characteristic information includes a portion of information training in gender, age, and liver index information of the sample patient. More specifically, the aamap model computes a corresponding first predictive score based on the gender, age, albumin, total bilirubin, and platelet count of the sample patient in liver index information, i.e., the first sample characteristic information includes gender, age, albumin, total bilirubin, and platelet count.
That is, the first model contains only three common clinical test indexes, namely Albumin (ALB), total Bilirubin (TBIL) and platelet count (PLT), except for age and gender, so that the risk prediction model and the device based on the risk prediction model can be well popularized in primary hospitals, and the aap score can obtain better evaluation performance in patients in liver inflammation activity period because the index of liver inflammation is not involved.
The prediction evaluation module 30 is configured to input sample evaluation information of a sample patient into the constructed first model, and obtain a first prediction score corresponding to the sample patient output by the first model.
The second construction module 40 is configured to determine second sample feature information from the first prediction score and the sample evaluation information of the sample patient, and construct a second model based on the second sample feature information and the first prediction score of the sample patient, where the second model is used to predict the second prediction score, and the second prediction score is also used to characterize the hepatocellular carcinoma monitoring condition corresponding to the sample patient.
In the embodiment of the invention, the second prediction score is obtained based on a second model, wherein the second model also adopts an aMAP score model, and the second model is an aMAP-2 model.
The second sample characteristic information includes part of information in liver index information of the sample patient. More specifically, the aap-2 model calculates a corresponding second predictive score based on the first predictive score of the sample patient, alpha fetoprotein in the liver index information, i.e., the second sample characteristic information includes the valley first predictive score and alpha fetoprotein.
A third construction module 50 is configured to sequentially process the constructed first model and the constructed second model to construct a third model. The third model is used for predicting a third prediction score, and the third prediction score is also used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
In the embodiment of the present invention, the third prediction score is obtained based on a third model, and since the third model is obtained based on a constructed first risk prediction model and a constructed second model which are sequentially applied, the third model is an aamap-2 model. Specifically, during actual application of the third model, sequential application of the first model and the second model means that a first prediction score of the to-be-evaluated person is obtained by using the first model, and when the first prediction score corresponding to the first follow-up visit of the to-be-evaluated person exceeds the first score, a second prediction score of the to-be-evaluated person is obtained by using the second model, and a third prediction score of the to-be-evaluated person is obtained according to the previous first prediction score and the second prediction score; and when the first predictive score corresponding to the to-be-evaluated person does not exceed the first score, obtaining a third predictive score directly according to the first predictive score.
Therefore, the third predictive score obtained based on the third model has higher hepatocellular carcinoma monitoring precision and good layering monitoring effect on hepatocellular carcinoma, and can guide doctors to develop better treatment schemes for patients.
According to the modeling device of the hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, the sample evaluation information of which the longitudinal data type comprises the serological markers is extracted from the longitudinal follow-up case information of a sample patient, the first model is constructed based on the sample evaluation information, the second model is constructed based on the sample evaluation information and the first prediction score output by the first model, and finally the final third model is obtained based on the first model and the second model by sequential application.
Fig. 6 illustrates a physical schematic diagram of an electronic device, as shown in fig. 6, which may include: processor 410, communication interface (Communications Interface) 420, memory 430 and communication bus 440, wherein processor 410, communication interface 420 and memory 430 communicate with each other via communication bus 440. The processor 410 may invoke logic commands in the memory 430 to perform a method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data, the method comprising:
longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information;
determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
Determining second sample feature information from the first predictive score and the sample assessment information for the sample patient, and constructing a second model based on the second sample feature information and the first predictive score for the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
sequentially processing the constructed first model and the constructed second model to construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
In addition, the logic commands in the memory 430 described above may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a separate medium. Based on this understanding, the technical solution of the present invention may be embodied in essence or a part contributing to the prior art or a part of the technical solution, or in the form of a software medium stored in a storage medium, including several commands for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program medium, where the computer program medium includes a computer program, where the computer program is stored on a non-transitory computer readable storage medium, and when the computer program is executed by a processor, the computer is capable of executing a modeling method of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data provided by the above methods, and the method includes:
longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information;
determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
determining second sample feature information from the first predictive score and the sample assessment information for the sample patient, and constructing a second model based on the second sample feature information and the first predictive score for the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
Sequentially processing the constructed first model and the constructed second model to construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform a method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data provided by the above methods, the method comprising:
longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information;
determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
Determining second sample feature information from the first predictive score and the sample assessment information for the sample patient, and constructing a second model based on the second sample feature information and the first predictive score for the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
sequentially processing the constructed first model and the constructed second model to construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the corresponding hepatocellular carcinoma monitoring condition of the sample patient.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in some part contributing to the prior art in the form of a software medium, which may be stored in a computer readable storage medium such as ROM/RAM, a magnetic disk, an optical disk, etc., including several commands for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. A method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, the method comprising:
longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information;
determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
determining second sample feature information from the first predictive score and the sample assessment information for the sample patient, and constructing a second model based on the second sample feature information and the first predictive score for the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
Sequentially processing the constructed first model and the constructed second model to construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient; when the third model is applied, the third prediction score is obtained through the following steps:
longitudinal follow-up case information of a person to be evaluated is obtained, the longitudinal follow-up case information is processed, and evaluation information of the person to be evaluated is extracted from the longitudinal follow-up case information;
determining first characteristic information of the first follow-up visit from the evaluation information of the first follow-up visit;
inputting the first characteristic information of the first follow-up visit into the first model to obtain the first prediction score corresponding to the first follow-up visit of the to-be-evaluated person output by the first model;
determining second characteristic information of the current follow-up visit from the first predictive score of the current follow-up visit and the evaluation information of the previous follow-up visit of the person to be evaluated when the first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds a first score;
inputting the second characteristic information of the current follow-up visit into the second model to obtain the second prediction score corresponding to the current follow-up visit of the evaluator output by the second model, and obtaining a third prediction score of the evaluator based on the first prediction score of the first follow-up visit and the second prediction score of the current follow-up visit;
And when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person is not more than the first score, obtaining a third predictive score of the to-be-evaluated person based on the first predictive score of the first follow-up visit.
2. The method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data according to claim 1, wherein the longitudinal follow-up case information of a sample patient is obtained, the longitudinal follow-up case information is processed, and sample evaluation information of the sample patient is extracted from the longitudinal follow-up case information, specifically comprising the following steps:
acquiring the longitudinal follow-up case information of the sample patient, and determining the type of the longitudinal follow-up case information;
if the longitudinal follow-up case information is determined to be unstructured type information, carrying out structuring treatment on the unstructured longitudinal follow-up case information to obtain structured longitudinal follow-up case information;
screening from the structured longitudinal follow-up case information, and extracting the sample evaluation information.
3. The method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data of claim 1, wherein the first model is modeled by:
Determining an initial evaluation factor corresponding to the first sample characteristic information;
adjusting the initial evaluation factor to obtain a first evaluation factor;
the first model is constructed based on the first sample characteristic information and the first evaluation factor.
4. A method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data as defined in claim 3, wherein the inputting the sample evaluation information of the sample patient into the constructed first model, obtaining a first prediction score corresponding to the sample patient output by the first model, specifically comprises:
inputting the sample evaluation information of a sample patient into the first model, and processing the sample evaluation information based on the first sample characteristic information and the first evaluation factor corresponding to the first model to obtain the first prediction score output by the first model.
5. A method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data in accordance with claim 3, wherein the second model is modeled by:
grouping the sample patients to obtain a sample patient of a diseased group and a sample patient of a non-diseased group;
Fitting the sample evaluation information of the sample patients in the diseased group and the non-diseased group and a first prediction score generated by a previous follow-up based on an initial longitudinal model to respectively obtain a first group average contour of the diseased group and a second group average contour of the non-diseased group; the initial longitudinal model adopts a multi-element linear mixed effect model;
the second model is constructed based on the first set of average contours and the second set of average contours.
6. The method of modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data of claim 5 wherein the second predictive score is obtained by:
when a first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds a first score, longitudinal follow-up visit case information of the person to be evaluated in the next follow-up visit and the previous follow-up visit and the first predictive score of each follow-up visit are obtained, the longitudinal follow-up visit case information is processed, and evaluation information of the person to be evaluated is extracted from the longitudinal follow-up visit case information;
determining second characteristic information from first prediction scores and evaluation information corresponding to multiple follow-up visits of the person to be evaluated;
Inputting the first predictive scores and the evaluation information of the multiple follow-up visits into the second model, and calculating a longitudinal profile corresponding to the person to be evaluated based on the initial longitudinal model;
comparing the longitudinal profile with the first group average profile and the second group average profile respectively to obtain a first grouping probability of the evaluators to be evaluated into the diseased group and a second grouping probability of the evaluators to be evaluated into the non-diseased group;
and calculating the second prediction score corresponding to the to-be-evaluated person based on the first grouping probability.
7. A modeling apparatus for a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data, the apparatus comprising:
the information extraction module is used for acquiring longitudinal follow-up case information of the sample patient, processing the longitudinal follow-up case information and extracting sample evaluation information of the sample patient from the longitudinal follow-up case information;
a first construction module for determining first sample characteristic information from the sample evaluation information of the sample patient and constructing a first model based on the first sample characteristic information of the sample patient; the first model is used for predicting a first prediction score, and the first prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
The prediction evaluation module is used for inputting the sample evaluation information of the sample patient into the constructed first model to obtain a first prediction score corresponding to the sample patient output by the first model;
a second construction module for determining second sample characteristic information from the first predictive score and the sample evaluation information of the sample patient, and constructing a second model based on the second sample characteristic information and the first predictive score of the sample patient; the second model is used for predicting a second prediction score, and the second prediction score is used for representing the hepatocellular carcinoma monitoring condition corresponding to the sample patient;
a third construction module, configured to sequentially process the constructed first model and the constructed second model, and construct a third model; the third model is used for predicting a third prediction score, and the third prediction score is used for representing a hepatocellular carcinoma monitoring condition corresponding to a sample patient, wherein when the third model is applied, the third prediction score is obtained through the following steps:
longitudinal follow-up case information of a person to be evaluated is obtained, the longitudinal follow-up case information is processed, and evaluation information of the person to be evaluated is extracted from the longitudinal follow-up case information;
Determining first characteristic information of the first follow-up visit from the evaluation information of the first follow-up visit;
inputting the first characteristic information of the first follow-up visit into the first model to obtain the first prediction score corresponding to the first follow-up visit of the to-be-evaluated person output by the first model;
determining second characteristic information of the current follow-up visit from the first predictive score of the current follow-up visit and the evaluation information of the previous follow-up visit of the person to be evaluated when the first predictive score corresponding to the first follow-up visit of the person to be evaluated exceeds a first score;
inputting the second characteristic information of the current follow-up visit into the second model to obtain the second prediction score corresponding to the current follow-up visit of the evaluator output by the second model, and obtaining a third prediction score of the evaluator based on the first prediction score of the first follow-up visit and the second prediction score of the current follow-up visit;
and when the first predictive score corresponding to the first follow-up visit of the to-be-evaluated person is not more than the first score, obtaining a third predictive score of the to-be-evaluated person based on the first predictive score of the first follow-up visit.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor, when executing the program, implements the steps of the method for modeling a hepatocellular carcinoma monitoring model based on longitudinal multi-dimensional data as defined in any one of claims 1 to 6.
9. A non-transitory computer readable storage medium, having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the modeling method of a hepatocellular carcinoma monitoring model based on longitudinal multidimensional data as claimed in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310154793.1A CN115862850B (en) | 2023-02-23 | 2023-02-23 | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310154793.1A CN115862850B (en) | 2023-02-23 | 2023-02-23 | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115862850A CN115862850A (en) | 2023-03-28 |
| CN115862850B true CN115862850B (en) | 2023-05-16 |
Family
ID=85658731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310154793.1A Active CN115862850B (en) | 2023-02-23 | 2023-02-23 | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115862850B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116167462B (en) * | 2023-04-23 | 2023-08-15 | 南方医科大学南方医院 | Hepatocellular carcinoma prediction model construction method, and hepatocellular carcinoma prediction method and device |
| CN116386901B (en) * | 2023-06-05 | 2024-01-30 | 南方医科大学南方医院 | Method and device for constructing clinical cure prediction model of hepatitis B, prediction method and device |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9779213B2 (en) * | 2008-07-25 | 2017-10-03 | Fundacao D. Anna Sommer Champalimaud E Dr. Carlos Montez Champalimaud | System for evaluating a pathological stage of prostate cancer |
| WO2011130495A1 (en) * | 2010-04-14 | 2011-10-20 | Nuvera Biosciences, Inc. | Methods of evaluating response to cancer therapy |
| WO2016040732A1 (en) * | 2014-09-11 | 2016-03-17 | The Regents Of The University Of Michigan | Machine learning for hepatitis c |
| TWI569012B (en) * | 2015-12-29 | 2017-02-01 | 財團法人工業技術研究院 | Method for estimating a risk for a subject suffering from hepatocellular carcinoma and method for prognosis of hepatocellular carcinoma |
| WO2019171049A1 (en) * | 2018-03-06 | 2019-09-12 | Ieso Digital Health Limited | Improvements in or relating to psychological profiles |
| WO2022221712A1 (en) * | 2021-04-15 | 2022-10-20 | Curemetrix, Inc. | Detecting, scoring and predicting disease risk using multiple medical-imaging modalities |
| CN115602319B (en) * | 2022-12-07 | 2023-03-10 | 南方医科大学南方医院 | Noninvasive hepatic fibrosis assessment device |
-
2023
- 2023-02-23 CN CN202310154793.1A patent/CN115862850B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115862850A (en) | 2023-03-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115862850B (en) | Modeling method and device of hepatocellular carcinoma monitoring model based on longitudinal multidimensional data | |
| US8793144B2 (en) | Treatment effect prediction system, a treatment effect prediction method, and a computer program product thereof | |
| Chauhan et al. | Heart disease prediction using evolutionary rule learning | |
| Zhou et al. | Multiple imputation in public health research | |
| EP3306500A1 (en) | Method for analysing medical treatment data based on deep learning, and intelligent analyser thereof | |
| US20160063212A1 (en) | System for Generating and Updating Treatment Guidelines and Estimating Effect Size of Treatment Steps | |
| AU2007240837A1 (en) | Personalized prognosis modeling in medical treatment planning | |
| Yan et al. | Selecting critical clinical features for heart diseases diagnosis with a real-coded genetic algorithm | |
| Luong et al. | A k-means approach to clustering disease progressions | |
| CN107436993A (en) | Establish the method and server of ICU conditions of patients assessment models | |
| Pang et al. | Hybrid outcome prediction model for severe traumatic brain injury | |
| CN113160986B (en) | Model construction method and system for predicting development of systemic inflammatory response syndrome | |
| Inácio et al. | Nonparametric Bayesian estimation of the three‐way receiver operating characteristic surface | |
| CN114023441A (en) | Severe AKI early risk assessment model and device based on interpretable machine learning model and development method thereof | |
| Belderrar et al. | Real-time estimation of hospital discharge using fuzzy radial basis function network and electronic health record data | |
| Wanriko et al. | Risk assessment of pregnancy-induced hypertension using a machine learning approach | |
| Karim et al. | Sistem Pakar dalam Mengidentifikasi Gejala Stroke Menggunakan Metode Naive Bayes | |
| CN112071431B (en) | Clinical path automatic generation method and system based on deep learning and knowledge graph | |
| MunishKhanna et al. | A novel approach for human diseases prediction using nature inspired computing & machine learning approach | |
| CN112967803A (en) | Early mortality prediction method and system for emergency patients based on integrated model | |
| Derevitskii et al. | Analysis course of the disease of type 2 diabetes patients using Markov chains and clustering methods. | |
| CN116564521A (en) | Chronic disease risk assessment model establishment method, medium and system | |
| CN116884559A (en) | Language model-based image report generation method and system | |
| Nazlı et al. | Classification of Coronary Artery Disease Using Different Machine Learning Algorithms | |
| Kumarage et al. | Efficient feature selection for prediction of diabetic using LASSO |
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 |