Stroke prognosis prediction method and system, computer equipment and storage medium

Abstract

The invention provides a stroke prognosis prediction method and system, computer equipment and a storage medium. The method comprises the following steps: acquiring to-be-analyzed feature data of a stroke patient; the to-be-analyzed feature data comprises age, gender, native place, mRS score during discharge, stroke-related medical history features and NIHSS score change features of hospital admission and discharge; inputting the to-be-analyzed data into a pre-constructed stroke prognosis prediction model for prognosis prediction to obtain a corresponding prognosis prediction result; the stroke prognosis prediction model is obtained by training a data set with the same features as the to-be-analyzed feature data. According to the method, the stroke prognosis prediction model is constructed by screening key data feature training which is high in obtainability and can be simultaneously used for accurate prediction of different post-stroke outcomes, so that the high efficiency and accuracy of post-stroke outcome prediction can be effectively improved, and the generalization and usability of the model are ensured; and reliable technical support is provided for health management of the stroke patient.

Description

Technical Field

[0001] The present invention relates to the technical field of stroke prognosis prediction, and in particular to a stroke prognosis prediction method, system, computer equipment and storage medium. Background Art

[0002] Stroke is characterized by high incidence, high disability, high mortality, high recurrence and limited treatment time window. It is the leading cause of death and disability in the world. Its treatment costs also impose a huge economic burden on the country, society and families. Timely and reliable stroke prognosis prediction is crucial for the health management of stroke patients.

[0003] Existing stroke prognosis prediction methods mainly include the establishment of post-stroke outcome prediction models based on traditional regression analysis methods and machine learning methods for predicting the outcomes of stroke patients. However, the existing post-stroke outcome prediction models have great application limitations and are difficult to promote and use effectively: 1) The construction of the model mainly depends on the imaging data or laboratory test results of stroke patients, and such indicators involve patient privacy and are poorly available in practical applications; 2) Different data features need to be used to construct different post-stroke outcome prediction models for different post-stroke outcome predictions, which cannot meet the convenience of using the same features to predict different stroke outcomes; 3) The area under the curve (ACU) predicted by the model is poor and cannot meet the needs of clinical promotion and application. Therefore, it is urgent to provide a high-reliability stroke prognosis prediction method that does not rely entirely on imaging data or laboratory test results and meets the common features of different post-stroke outcome predictions. Summary of the invention

[0004] The purpose of the present invention is to provide a stroke prognosis prediction method. By studying and analyzing the stroke research data set, key data features that are independent of imaging data or laboratory test results, have strong data accessibility, and can be simultaneously applicable to the accurate prediction of different post-stroke outcomes are screened out. Based on the key data feature training, an efficient, accurate, and risk identification and discrimination stroke prognosis prediction model is constructed for use in predicting the prognosis outcomes of stroke patients. While effectively improving the efficiency and accuracy of post-stroke outcome prediction, the generalization and ease of use of the prediction model can be ensured, which is convenient for large-scale promotion and use, thereby providing reliable technical support for the health management of stroke patients.

[0005] In order to achieve the above objectives, it is necessary to provide a stroke prognosis prediction method, system, computer device and storage medium in response to the above technical problems.

[0006] In a first aspect, an embodiment of the present invention provides a method for predicting stroke prognosis, the method comprising the following steps:

[0007] Obtaining the characteristic data to be analyzed of the stroke patients; the characteristic data to be analyzed include age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics at admission and discharge;

[0008] The data to be analyzed are input into a pre-built stroke prognosis prediction model for prognosis prediction to obtain corresponding prognosis prediction results; the stroke prognosis prediction model is trained based on a data set with the same characteristics as the characteristic data to be analyzed.

[0009] Furthermore, the stroke-related medical history characteristics include the presence or absence of a history of transient ischemic attack, the presence or absence of a history of diabetes, and the presence or absence of a history of transient ischemic attack; the admission and discharge NIHSS score change characteristics include the admission and discharge NIHSS total score change value and the corresponding admission and discharge NIHSS single score change values.

[0010] Furthermore, the steps of constructing the stroke prognosis prediction model include:

[0011] Acquire a stroke research data set; the stroke research data set includes a first stroke patient data set and a second stroke patient data set; the first stroke patient data set and the second stroke patient data set both include age, gender, native place, mRS score at discharge, stroke-related medical history characteristics, admission NIHSS total score and several corresponding admission NIHSS single-item scores, discharge NIHSS total score and several corresponding discharge NIHSS single-item scores;

[0012] After preprocessing the stroke research data set, a first training set, a first external validation set, and a second external validation set are obtained;

[0013] Extracting key features for stroke prognosis prediction from data features of the first training set according to the first training set and the XGBoost model;

[0014] Obtaining a corresponding second training set, a first validation set, and a second validation set according to the key features for stroke prognosis prediction and the corresponding prognostic outcomes in the first training set, the first external validation set, and the second external validation set, respectively;

[0015] The XGBoost model is trained according to the second training set, and the optimal parameter model is obtained through K-fold cross validation to obtain the stroke prognosis prediction model. The stroke prognosis prediction model is verified by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain a prediction model verification result.

[0016] Furthermore, the step of obtaining a first training set, a first external validation set, and a second external validation set after preprocessing the stroke patient data in the stroke research dataset includes:

[0017] The discharge NIHSS total score in each stroke patient data in the stroke research data set is subtracted from the admission NIHSS total score to obtain the corresponding admission and discharge NIHSS total score change value, and the corresponding discharge NIHSS single item score is subtracted from the admission NIHSS single item score to obtain the corresponding admission and discharge NIHSS single item score change value;

[0018] Adding the admission and discharge NIHSS total score change value and the corresponding admission and discharge NIHSS single score change value of each stroke patient data in the stroke research data set to the stroke research data set to obtain an updated stroke research data set;

[0019] The first stroke patient data set portion in the updated stroke research data set is divided into a first training set and a first external validation set, and a preset number of patient data are randomly selected from the second stroke patient data set portion in the updated stroke research data set to obtain a second external validation set.

[0020] Furthermore, the step of extracting key features for stroke prognosis prediction from the data features of the first training set according to the first training set and the XGBoost model includes:

[0021] Training the XGBoost model according to the first training set, and after the training is completed, obtaining a corresponding feature importance ranking table;

[0022] According to the feature importance ranking table, the first important features for predicting stroke prognosis are obtained; the first important features for predicting stroke prognosis include the change characteristics of NIHSS scores at admission and discharge and the mRS score at discharge;

[0023] Based on the feature accessibility suggestions of clinical analysis, the second important features for stroke prognosis prediction are obtained from the data features of the first training set; the second important features for stroke prognosis prediction include age, gender, place of origin and stroke-related medical history characteristics;

[0024] The first important feature for predicting prognosis of stroke and the second important feature for predicting prognosis of stroke are combined to obtain the key feature for predicting prognosis of stroke.

[0025] Furthermore, the step of training the XGBoost model according to the second training set, obtaining the optimal parameter model through K-fold cross validation, obtaining the stroke prognosis prediction model, and verifying the stroke prognosis prediction model by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain the prediction model verification result includes:

[0026] The XGBoost model is trained according to the training set and the inner and outer K-fold cross validation method, and the optimal parameters of the XGBoost model are obtained by grid search to obtain a stroke prognosis prediction model corresponding to the optimal parameters; the optimal parameters include a learning rate and a maximum depth;

[0027] Performing homologous data verification on the stroke prognosis prediction model according to the first verification set to obtain a corresponding model homologous data verification result;

[0028] The stroke prognosis prediction model is verified using non-homologous data according to the second verification set to obtain corresponding model heterogeneous data verification results.

[0029] Furthermore, the method further comprises:

[0030] The SHAP attribution analysis method is used to obtain the feature importance ranking corresponding to the prognosis prediction result, and the corresponding rehabilitation exercise guidance suggestions are generated according to the feature importance ranking.

[0031] In a second aspect, an embodiment of the present invention provides a stroke prognosis prediction system, the system comprising:

[0032] A data acquisition module is used to acquire characteristic data to be analyzed of stroke patients; the characteristic data to be analyzed includes age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics at admission and discharge;

[0033] The prognosis prediction module is used to input the data to be analyzed into a pre-built stroke prognosis prediction model for prognosis prediction to obtain a corresponding prognosis prediction result; the stroke prognosis prediction model is trained based on a data set with the same characteristics as the characteristic data to be analyzed.

[0034] In a third aspect, an embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0035] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0036] The present invention provides a method, system, computer device and storage medium for predicting the prognosis of a stroke. The method can obtain the characteristic data to be analyzed, including the age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics and NIHSS score change characteristics of the stroke patient, and input the data to be analyzed into a stroke prognosis prediction model pre-trained based on a data set with the same characteristics as the characteristic data to be analyzed for prognosis prediction, and obtain the corresponding technical solution for the prognosis prediction result. Compared with the prior art, the stroke prognosis prediction method screens out key data features that are independent of imaging data or laboratory test results, have strong data accessibility and can be simultaneously applied to the accurate prediction of different post-stroke outcomes through the research and analysis of the stroke research data set, and constructs an efficient, accurate and high risk identification and differentiation stroke prognosis prediction model based on the key data feature training for the prediction of the prognosis outcome of stroke patients. It can not only effectively improve the efficiency and accuracy of the prediction of the post-stroke outcome, but also ensure the generalization and ease of use of the prediction model, which is convenient for large-scale promotion and use, and thus provides reliable technical support for the health management of stroke patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a flow chart of a stroke prognosis prediction method in an embodiment of the present invention;

[0038] Figure 2 is a schematic diagram of feature data to be analyzed in an embodiment of the present invention;

[0039] Figure 3 1 is a schematic diagram of the feature importance ranking for screening the first important feature for predicting stroke prognosis in an embodiment of the present invention;

[0040] Figure 4 It is a schematic diagram of the comparison of the prediction results of the DISCO model in the embodiment of the present invention and the RRE model commonly used in the existing clinical practice;

[0041] Figure 5 It is a schematic diagram of the validation of the DISCO model for the 3-month recurrence risk stratification effect on the second training set, the first validation set, and the second validation set in an embodiment of the present invention;

[0042] Figure 6 is another schematic flow chart of a stroke prognosis prediction method in an embodiment of the present invention;

[0043] Figure 7 is a schematic diagram of the feature importance ranking result of the DISCO model obtained by the SHAP attribution analysis method in an embodiment of the present invention;

[0044] Figure 8 is a schematic structural diagram of a stroke prognosis prediction system in an embodiment of the present invention;

[0045] Fig. 9 is another structural schematic diagram of a stroke prognosis prediction system in an embodiment of the present invention;

[0046] Fig.10 It is a diagram of the internal structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and beneficial effects of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. Obviously, the embodiments described below are part of the embodiments of the present invention and are only used to illustrate the present invention, but are not used to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In one embodiment, Figure 1 As shown, a method for predicting stroke prognosis is provided, comprising the following steps:

[0049] S11. Obtain the characteristic data to be analyzed of the stroke patient; wherein, the characteristic data to be analyzed can be understood as the discharged patient data indicators that can be used to analyze different prognostic outcomes at the same time, which are obtained based on the key features of stroke prognosis prediction obtained by analyzing the existing stroke patient research data set and based on the relevant report information of the stroke patient at the time of discharge and admission; preferably, in this embodiment, the characteristic data to be analyzed include age, gender, place of origin, mRS (modified RankinScore) score at discharge, stroke-related medical history characteristics, and admission and discharge NIHSS score change characteristics. Among them, the stroke-related medical history characteristics can be understood as the relevant medical history data that may affect the development of stroke. Specifically, the stroke-related medical history characteristics include whether there is a history of transient ischemic attack, whether there is a history of diabetes, and whether there is a history of transient ischemic attack. In practical applications, the value of this type of indicator is 0 or 1, 0 means no, 1 means yes, and can be obtained based on the patient's complaint or filled in registration information.

[0050] The admission and discharge NIHSS (National Institutes of Health Stroke Scale) score change characteristics in this embodiment can be understood as the data change information obtained by comparing and analyzing the NIHSS score sheets (National Institutes of Health Stroke Scale) obtained when the patient is discharged and when he is admitted, that is, by calculating the difference between the data at discharge and the data at admission; specifically, the admission and discharge NIHSS score change characteristics include the admission and discharge NIHSS total score change value and a number of corresponding admission and discharge NIHSS single score change values, and the number of admission and discharge NIHSS single score change values ​​can be selected according to actual application requirements. In order to ensure the comprehensiveness of the analysis, Figure 2 As shown, it can include the single score change values ​​of 15 aspects in the NIHSS scale, including consciousness, speech, visual field, facial movement, upper limb movement, lower limb movement, coordinated movement, sensation, language comprehension and reading.

[0051] S12, input the data to be analyzed into the pre-built stroke prognosis prediction model for prognosis prediction, and obtain the corresponding prognosis prediction result; wherein, the construction of the stroke prognosis prediction model can be obtained in principle by using the existing machine learning model construction method that can realize prognosis prediction analysis, but in order to ensure the efficiency, accuracy and generalization of the prediction analysis, this embodiment preferably adopts the XGBoost algorithm to first analyze and screen the key features of stroke prognosis prediction for multiple existing classic stroke research data sets, and then train and verify the XGBoost model based on the key features of prognosis prediction of different data sets to obtain the final required stroke prognosis prediction model. Specifically, the steps of constructing the stroke prognosis prediction model include:

[0052] A stroke research dataset is obtained; the stroke research dataset includes a first stroke patient dataset and a second stroke patient dataset; the first stroke patient dataset and the second stroke patient dataset both include age, gender, native place, mRS score at discharge, stroke-related medical history characteristics, admission NIHSS total score and several corresponding admission NIHSS single scores, discharge NIHSS total score and several corresponding discharge NIHSS single scores. In order to ensure the reliability of the research data, the first stroke patient data set in this embodiment is the data of ischemic stroke patients after excluding the data of transient ischemic attack (TIA) patients from the China National Stroke Registry Cohort III (CNSR-III, a total of 15,166 patients were included, which is a national stroke or TIA registration cohort involving 201 hospitals in China, and the registration time is from August 2015 to March 22, 2018), and the second stroke patient data set is the data of ischemic stroke patients after excluding the data of TIA patients from the cohort data of Clopidogrel Combined with Aspirin for Acute Minor Stroke or Transient Ischemic Attack-II (CHANCE-2). It should be noted that the admission NIHSS single score and the discharge NIHSS single score in the first stroke patient data set and the second stroke patient data set are consistent with the NIHSS single score in the aforementioned feature data to be analyzed.

[0053] After preprocessing the stroke research data set, a first training set, a first external validation set, and a second external validation set are obtained; wherein the preprocessing mainly includes comparing and analyzing the total NIHSS score at admission and the corresponding single NIHSS scores at admission, the total NIHSS score at discharge and the corresponding single NIHSS scores at discharge for each patient, obtaining the change value of the total NIHSS score at admission and discharge and the corresponding change values ​​of the single NIHSS scores at admission and discharge, which are added to the current data set to obtain a new data set; specifically, the steps of obtaining the training set, the first external validation set, and the second external validation set after preprocessing the stroke patient data in the stroke research data set include:

[0054] The discharge NIHSS total score in each stroke patient data in the stroke research data set is subtracted from the admission NIHSS total score to obtain the corresponding admission and discharge NIHSS total score change value, and the corresponding discharge NIHSS single item score is subtracted from the admission NIHSS single item score to obtain the corresponding admission and discharge NIHSS single item score change value;

[0055] The admission and discharge NIHSS total score change value and the corresponding admission and discharge NIHSS single score change value of each stroke patient data in the stroke research dataset are added to the stroke research dataset to obtain an updated stroke research dataset, wherein the updated stroke research dataset can be understood as adding the admission and discharge NIHSS total score change value and the admission and discharge NIHSS single score change value obtained by the above analysis to the original stroke research dataset.

[0056] The first stroke patient data set portion in the updated stroke research data set is divided into a first training set and a first external validation set, and a preset number of patient data are randomly selected from the second stroke patient data set portion in the updated stroke research data set to obtain a second external validation set; wherein, the division ratio of the first training set and the first external validation set of the first stroke patient data set can be set according to actual application requirements. For example, if the first stroke patient data set is an ischemic stroke patient data set after the TIA (transient ischemic attack) patient data is removed from the China National Stroke Registry Cohort III (CNSR-III) data as mentioned above, 11,813 patients from 169 hospitals in the CNSR-III cohort can be randomly selected to be assigned to the first training set, and 2,127 patients from another 40 independent hospitals can be assigned to the first external validation set. It should be noted that the preset number used to construct the second external validation set can be determined according to actual application requirements and is not specifically limited here. For example, if the second stroke patient data set mentioned above is an ischemic stroke patient data set after excluding TIA patient data from the cohort data of the Clopidogrel Combined with Aspirin Treatment for Acute Minor Stroke or Transient Ischemic Attack-II (CHANCE-2) study, 5158 ischemic stroke patients can be selected as research subjects.

[0057] According to the first training set and the XGBoost model, key features for stroke prognosis prediction are extracted from the data features of the first training set; wherein the key features for stroke prognosis prediction can be understood as data features that have a greater impact on the outcome after stroke; specifically, according to the first training set and the XGBoost model, the step of extracting key features for stroke prognosis prediction from the data features of the first training set includes:

[0058] The XGBoost model is trained according to the first training set, and after the training is completed, a corresponding feature importance ranking table is obtained; wherein the feature importance ranking table can be obtained by the feature importance analysis and ranking function of the XGBoost algorithm after the XGBoost model is trained based on the first training set, or by analyzing the prediction contribution of each data feature in the first training set to the XGBoost model based on the SHAP attribution analysis method, or by comprehensively analyzing the analysis results of the two to obtain indicators with relatively high importance rankings in the results obtained by the two analysis methods, which are not specifically limited here.

[0059] According to the feature importance ranking table, the first important features for stroke prognosis prediction are obtained; wherein, the first important features for stroke prognosis prediction can be understood as data features that are ranked relatively high in the feature importance ranking table. The specific number of features can be determined in principle according to actual application requirements. In order to improve the convenience and efficiency of data acquisition while ensuring the reliability of subsequent model training, this embodiment preferably selects two data features with particularly prominent importance, such as Figure 3 As shown in the figure, the characteristics of the changes in the NIHSS scores upon admission and discharge and the mRS scores at discharge are ranked in the top two in the feature importance analysis and ranking function of the XGBoost algorithm and the feature importance ranking table obtained based on the SHAP attribution analysis method, and the corresponding scores are very high. It can be determined that the first important features for predicting stroke prognosis include the characteristics of the changes in the NIHSS scores upon admission and discharge and the mRS scores at discharge.

[0060] Based on the feature accessibility suggestions of clinical analysis, the second important features for stroke prognosis prediction are obtained from the data features of the first training set; wherein, the feature accessibility suggestions of clinical analysis can be understood as the results of investigation and consultation with relevant clinical experts on the influencing indicators of post-stroke outcomes; similarly, based on the convenience and efficiency of data acquisition, this embodiment preferably sets the second important features for stroke prognosis prediction to include age, gender, native place (the patient's geographical location in China) and stroke-related medical history characteristics, and the stroke-related medical history characteristics include the presence or absence of a history of transient ischemic attack (patient TOAST classification), the presence or absence of a history of diabetes, and the presence or absence of a history of transient ischemic attack (TIA history), etc.

[0061] The first important feature for predicting stroke prognosis and the second important feature for predicting stroke prognosis are combined to obtain the key features for predicting stroke prognosis; wherein the key features for predicting stroke prognosis preferably include age, gender, place of origin, mRS (modified Rankin Score) score at discharge, history of transient ischemic attack (patient TOAST classification), history of diabetes and history of transient ischemic attack (TIA history), change in total NIHSS score at admission and discharge, and change in the corresponding 15 individual NIHSS scores at admission and discharge.

[0062] According to the key features for predicting stroke prognosis and the corresponding prognostic outcomes in the first training set, the first external validation set and the second external validation set, the corresponding second training set, the first validation set and the second validation set are obtained respectively; that is, the second training set, the first validation set and the second validation set are data sets constructed by the key features for predicting stroke prognosis and the corresponding prognostic outcomes screened out in the first training set, the first external validation set and the second external validation set, respectively.

[0063] The XGBoost model is trained according to the second training set, and the optimal parameter model is obtained through K-fold cross validation to obtain the stroke prognosis prediction model, and the stroke prognosis prediction model is verified by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain a prediction model verification result; specifically, the XGBoost model is trained according to the second training set, and the optimal parameter model is obtained through K-fold cross validation to obtain the stroke prognosis prediction model, and the stroke prognosis prediction model is verified by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain a prediction model verification result. The steps include:

[0064] According to the second training set and the internal and external K-fold cross-validation method, the XGBoost model is trained, and the optimal parameters of the XGBoost model are obtained through grid search to obtain a stroke prognosis prediction model corresponding to the optimal parameters; the optimal parameters include learning rate and maximum depth; wherein, the specific training data and validation data division ratio used in the K-fold cross-validation method can be set according to actual application requirements, and this embodiment is preferably set to ten-fold cross-validation.

[0065] In order to ensure the efficiency and reliability of model training, when XGB Classifier was used for model fitting training based on the second training set and K-fold cross validation method, the two hyperparameters of learning rate (eta) and maximum depth (max depth) were adjusted by grid search, and the precision, recall rate, F1 score (the harmonic mean of precision and recall rate, F1 score = [(2*precision*recall rate) / (precision+recall rate)] and AUC value of the ROC curve (the area under the receiver operating curve, if the area under the curve is 0.5, it means random judgment, the larger the area under the curve, the better, and the ideal value is 1) of the model without parameter combination were obtained. The optimal parameters were screened out according to the AUC value of the ROC curve, and the XGBoost model corresponding to the optimal parameters was used as the initial prognosis prediction model.

[0066] The stroke prognosis prediction model is verified by homologous data according to the first verification set to obtain corresponding model homologous data verification results; wherein, the model homologous data verification results can be understood as the indicator value of the prediction effect of the verification model obtained based on the first verification set (the same source as the training set of the initial prognosis prediction model), and the indicator can be the prediction mean square error (MSE) of the model.

[0067] The stroke prognosis prediction model is verified by non-homologous data according to the second verification set to obtain corresponding model heterogeneous data verification results; wherein, the model heterogeneous data verification results can be understood as the index value of the prediction effect of the verification model obtained based on the second verification set (different from the training set source of the initial prognosis prediction model), which is consistent with the index selected for the above-mentioned model homologous data verification results, and the model's prediction mean square error (MSE) can also be selected.

[0068] In addition, considering that in actual applications, the stroke prognosis prediction model trained based on the second training set may perform poorly on the validation set, in order to improve the generalization ability of the prediction model as much as possible, this embodiment preferably fine-tunes the parameters of the stroke prognosis prediction model according to the model homologous data verification results and the model heterologous data verification results; specifically, the process of fine-tuning the parameters of the stroke prognosis prediction model includes: comparing the obtained model homologous data verification results and the model heterologous data verification results with the corresponding homologous verification index thresholds and heterologous verification index thresholds, respectively; if the model homologous data verification results are If both the verification result of the model homologous data and the verification result of the model heterogeneous data are less than the corresponding index threshold, it is considered that the generalization ability of the stroke prognosis prediction model obtained above is guaranteed, and it can be directly used as the final stroke prognosis prediction model. On the contrary, if any one of the model homologous data verification result and the model heterogeneous data verification result is greater than the corresponding index threshold, it is considered that the generalization ability of the stroke prognosis prediction model obtained above needs to be further improved. According to the second training set, the model parameters of the stroke prognosis prediction model are optimized and trained based on the preset loss function to obtain the final stroke prognosis prediction model; wherein, the preset loss function is expressed as:

[0069]

[0070] Where L represents the model prediction loss value; M1 and M2 represent the training sample subset and validation sample subset divided in the K-fold cross training in the second training set; n1 and n2 represent the number of samples in the training sample subset and validation sample subset respectively; y i，ture and i，pre Respectively represent the true value and predicted value of the i-th sample in the training sample subset M1; y j，ture and j，pre They represent the true value and predicted value of the jth sample in the validation sample subset respectively; λ represents the weight coefficient.

[0071] Through the above method steps, a stroke prognosis prediction model with high accuracy and risk discrimination can be obtained when verified on different data sources. For the convenience of subsequent description, the obtained stroke prognosis prediction model is named DISCO (Delta-NIHSS based post-Stroke Composite Outcome) model.

[0072] It should be noted that the stroke outcome labels used in the construction of the above stroke prognosis prediction model vary depending on the prediction requirements: when predicting stroke recurrence, the corresponding recurrence index value is used as the training label; when predicting stroke disability, the corresponding mRS score variation index (mRS>2 is 1, mRS<=2 is 0) is used as the training label; when predicting death after stroke, the corresponding death index value is used as the training label. The construction of prediction models for different prognostic outcomes can be obtained through the above method steps. If the obtained stroke prognosis prediction model is for the prediction of stroke recurrence outcome, the corresponding prediction result is the stroke recurrence probability value, if the obtained stroke prognosis prediction model is for the prediction of stroke death outcome, the corresponding prediction result is the stroke death probability value, if the obtained stroke prognosis prediction model is for the prediction of stroke disability outcome, the corresponding prediction result is the probability value of mRS>2.

[0073] The embodiment of the present invention provides a technical solution for obtaining feature data to be analyzed of stroke patients, including age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics between admission and discharge, and inputting the data to be analyzed into a stroke prognosis prediction model pre-trained based on a data set having the same characteristics as the feature data to be analyzed to perform prognosis prediction to obtain corresponding prognosis prediction results. By studying and analyzing the stroke research data set, key data features that are independent of imaging data or laboratory test results, have strong data accessibility, and can be simultaneously applied to the accurate prediction of different post-stroke outcomes are screened out. Based on the key data feature training, an efficient, accurate, and risk-identifying high-discrimination stroke prognosis prediction model is constructed for use in predicting the prognosis outcomes of stroke patients. This not only effectively improves the efficiency and accuracy of post-stroke outcome prediction, but also ensures the generalization and ease of use of the prediction model, facilitates large-scale promotion and use, and thus provides reliable technical support for the health management of stroke patients.

[0074] In order to verify the application effect of the stroke prognosis prediction method provided by the present invention, this embodiment also obtains a first training set, a first external validation set, and a second external validation set based on the first stroke patient dataset constructed based on the China National Stroke Registry Cohort III (CNSR-III) data in the above embodiment and a second stroke patient dataset constructed based on the efficacy study of clopidogrel combined with aspirin in the treatment of acute minor stroke or transient ischemic attack-II (CHANCE-2) cohort data. Taking the 3-month recurrence risk prediction of 3-month stroke patients as an example, the DISCO model is compared and analyzed with the existing clinically commonly used RRE model, and the obtained Figure 4 The analysis results shown in Figure 4 As shown, the DISCO model provided by the present invention has better prediction effect.

[0075] In addition, the prediction effect of the DISCO model on the three outcomes of recurrence, death and disability within 3 months was verified based on the second training set, the first validation set and the second validation set, and the model prediction index results were obtained as shown in Tables 1-3; at the same time, the DISCO model was also verified on the second training set, the first validation set and the second validation set for the three-month recurrence risk stratification effect, and the results were as follows Figure 5 Results shown.

[0076] Table 1. Prediction results of DISCO model on the second training set

[0077]

[0078] Table 2 Prediction results of DISCO model on the first validation set

[0079]

[0080] Table 3 Prediction results of DISCO model on the second validation set

[0081]

[0082]

[0083] Based on the performance of the DISCO model on the second training set in Table 1, it can be seen that the DISCO model has good prediction results for the 3-month stroke recurrence event that is of greatest clinical concern, with AUC=0.80, precision=0.98, recall=0.93, and F1=0.95. The prediction effect for death within 3 months is better than that for recurrence, with AUC>0.8, precision>0.99, recall>0.8, and F1>0.89. Although the prediction of 3-month disability mRS is relatively weaker than recurrence, the AUC value can still be guaranteed to be between 0.7 and 0.8.

[0084] Similarly, based on the performance of the DISCO model on the first validation set in Table 2, it can be seen that DISCO has an AUC of 0.74, precision of 0.96, recall of 0.93, and F1 of 0.94 for 3-month stroke recurrence events.

[0085] Based on the performance of the DISCO model on the second validation set in Table 3, it can be seen that the DISCO model has an AUC of 0.80, precision of 0.96, recall of 0.93, and F1 of 0.95 for the prediction of 3-month stroke recurrence events. In addition, in both validation sets, the DISCO model also has a better prediction effect on death within 3 months than recurrence, with AUC>0.8, precision>0.98, recall>0.80, and F1>0.88. Although the prediction of 3-month disability mRS is relatively weaker than recurrence, it can also ensure that the AUC value is around 0.7.

[0086] like Figure 5 As shown in the figure, the model calculated the risk score of stroke for each person within 3 months. The relative risks (RR) of the highest risk group and the reference group were 33.6, 27.46 and 46.3, respectively. This evaluation result shows that the DISCO model has a very good ability to identify people at high risk of stroke recurrence within 3 months.

[0087] In addition, in order to further enhance the application value of the above-mentioned stroke prognosis prediction model, the embodiment of the present invention preferably provides personalized rehabilitation exercise guidance suggestions based on the importance ranking of each patient's specific feature data and the professional suggestions of clinical workers on the basis of predicting the required prognosis outcome based on the feature data to be analyzed provided by discharged stroke patients. Specifically, Figure 6 Said method further comprises:

[0088] S13. Obtain the feature importance ranking corresponding to the prognosis prediction result through SHAP attribution analysis, and generate corresponding rehabilitation training guidance suggestions based on the feature importance ranking; wherein, the feature importance ranking can be understood as the ranking of the contribution of each feature data to be analyzed provided by the stroke patient at the time of discharge to the predicted outcome. Figure 7 After obtaining the feature importance ranking results of the DISCO model shown in the figure, we can use the features with the highest contribution ranking as a reference and combine the experience of clinical workers to give targeted rehabilitation exercise guidance suggestions. We can also obtain it by comparing the actual feature importance ranking with the corresponding guidance suggestion knowledge base pre-built based on different feature ranking results. In order to further improve the scientificity and intelligence of the generation of rehabilitation exercise guidance suggestions, we can also construct a reinforcement learning model with the feature importance ranking results as state variables and the rehabilitation exercise guidance suggestions as corresponding action instructions. In actual applications, the obtained feature importance ranking can be input into the trained reinforcement learning model to obtain the required rehabilitation exercise guidance suggestions.

[0089] The embodiment of the present invention obtains the feature data to be analyzed including age, gender, native place, mRS score at discharge, stroke-related medical history characteristics and NIHSS score change characteristics between admission and discharge of the stroke patient, inputs the data to be analyzed into a stroke prognosis prediction model pre-trained based on a data set with the same characteristics as the feature data to be analyzed to perform prognosis prediction, and obtains the corresponding prognosis prediction result, and obtains the feature importance ranking corresponding to the prognosis prediction result through the SHAP attribution analysis method, and generates corresponding rehabilitation guidance suggestions according to the feature importance ranking. The technical solution can not only effectively improve the efficiency and accuracy of post-stroke outcome prediction, but also ensure the generalization and ease of use of the prediction model, facilitate large-scale promotion and use, and can also give corresponding rehabilitation guidance suggestions based on the prediction results, which has high health management guidance significance, and thus provides reliable technical support for the health management of stroke patients.

[0090] It should be noted that although the steps in the above flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders.

[0091] In one embodiment, Figure 8 As shown, a stroke prognosis prediction system is provided, the system comprising:

[0092] Data acquisition module 1 is used to acquire characteristic data to be analyzed of stroke patients; the characteristic data to be analyzed includes age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics at admission and discharge;

[0093] The prognosis prediction module 2 is used to input the data to be analyzed into a pre-built stroke prognosis prediction model for prognosis prediction to obtain corresponding prognosis prediction results.

[0094] In one embodiment, Fig. 9 As shown, a stroke prognosis prediction system is provided, and the system also includes:

[0095] The suggestion generation module 3 is used to obtain the feature importance ranking corresponding to the prognosis prediction result through the SHAP attribution analysis method, and generate corresponding rehabilitation guidance suggestions according to the feature importance ranking.

[0096] For the specific limitations of the stroke prognosis prediction system, please refer to the limitations of the stroke prognosis prediction method above, and the corresponding technical effects can also be equivalently obtained, which will not be repeated here. Each module in the above-mentioned stroke prognosis prediction system can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0097] Fig.10 FIG. 1 shows an internal structure diagram of a computer device in an embodiment, and the computer device may specifically be a terminal or a server. Fig.10 As shown, the computer device includes a processor, a memory, a network interface, a display, a camera and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a stroke prognosis prediction method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0098] It can be understood by those skilled in the art that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computing device may include more or less components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0099] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the computer program.

[0100] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0101] In summary, the embodiments of the present invention provide a stroke prognosis prediction method, system, computer device and storage medium. The stroke prognosis prediction method realizes obtaining the feature data to be analyzed including age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics and NIHSS score change characteristics of stroke patients, inputting the data to be analyzed into a stroke prognosis prediction model pre-trained based on a data set with the same characteristics as the feature data to be analyzed to perform prognosis prediction, and obtain corresponding prognosis prediction results, and obtain the feature importance ranking corresponding to the prognosis prediction results through the SHAP attribution analysis method, and generate corresponding rehabilitation guidance technical solutions according to the feature importance ranking suggestions. This method can not only effectively improve the efficiency and accuracy of post-stroke outcome prediction, but also ensure the generalization and ease of use of the prediction model, facilitate large-scale promotion and use, and can also give corresponding rehabilitation guidance suggestions based on the prediction results, which has high health management guidance significance, and thus provides reliable technical support for the health management of stroke patients.

[0102] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0103] The above-mentioned embodiments only express several preferred implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principle of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for predicting stroke prognosis, characterized in that: The method comprises the following steps: Obtaining the characteristic data to be analyzed of the stroke patients; the characteristic data to be analyzed include age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics at admission and discharge; The data to be analyzed are input into a pre-built stroke prognosis prediction model for prognosis prediction to obtain corresponding prognosis prediction results; the stroke prognosis prediction model is trained based on a data set with the same characteristics as the characteristic data to be analyzed.

2. The method for predicting stroke prognosis according to claim 1, characterized in that: The stroke-related medical history characteristics include the presence or absence of a history of transient ischemic attack, the presence or absence of a history of diabetes, and the presence or absence of a history of transient ischemic attack; the NIHSS score change characteristics upon admission and discharge include the change value of the NIHSS total score upon admission and discharge and the corresponding change values ​​of several NIHSS single score upon admission and discharge.

3. The method for predicting stroke prognosis according to claim 1, characterized in that: The steps of constructing the stroke prognosis prediction model include: Acquire a stroke research data set; the stroke research data set includes a first stroke patient data set and a second stroke patient data set; the first stroke patient data set and the second stroke patient data set both include age, gender, native place, mRS score at discharge, stroke-related medical history characteristics, admission NIHSS total score and several corresponding admission NIHSS single-item scores, discharge NIHSS total score and several corresponding discharge NIHSS single-item scores; After preprocessing the stroke research data set, a first training set, a first external validation set, and a second external validation set are obtained; Extracting key features for stroke prognosis prediction from data features of the first training set according to the first training set and the XGBoost model; Obtaining a corresponding second training set, a first validation set, and a second validation set according to the key features for stroke prognosis prediction and the corresponding prognostic outcomes in the first training set, the first external validation set, and the second external validation set, respectively; The XGBoost model is trained according to the second training set, and the optimal parameter model is obtained through K-fold cross validation to obtain the stroke prognosis prediction model. The stroke prognosis prediction model is verified by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain a prediction model verification result.

4. The method for predicting stroke prognosis according to claim 3, characterized in that: The step of obtaining a first training set, a first external validation set, and a second external validation set after preprocessing the stroke patient data in the stroke research data set comprises: The discharge NIHSS total score in each stroke patient data in the stroke research data set is subtracted from the admission NIHSS total score to obtain the corresponding admission and discharge NIHSS total score change value, and the corresponding discharge NIHSS single item score is subtracted from the admission NIHSS single item score to obtain the corresponding admission and discharge NIHSS single item score change value; Adding the admission and discharge NIHSS total score change value and the corresponding admission and discharge NIHSS single score change value of each stroke patient data in the stroke research data set to the stroke research data set to obtain an updated stroke research data set; The first stroke patient data set portion in the updated stroke research data set is divided into a training set and a first external validation set, and a preset number of patient data are randomly selected from the second stroke patient data set portion in the updated stroke research data set to obtain a second external validation set.

5. The method for predicting stroke prognosis according to claim 3, characterized in that: The step of extracting key features for stroke prognosis prediction from the data features of the first training set according to the first training set and the XGBoost model comprises: Training the XGBoost model according to the first training set, and after the training is completed, obtaining a corresponding feature importance ranking table; According to the feature importance ranking table, the first important features for predicting stroke prognosis are obtained; the first important features for predicting stroke prognosis include the change characteristics of NIHSS scores at admission and discharge and the mRS score at discharge; Based on the feature accessibility suggestions of clinical analysis, the second important features for stroke prognosis prediction are obtained from the data features of the first training set; the second important features for stroke prognosis prediction include age, gender, place of origin and stroke-related medical history characteristics; The first important feature for predicting prognosis of stroke and the second important feature for predicting prognosis of stroke are combined to obtain the key feature for predicting prognosis of stroke.

6. The method for predicting stroke prognosis according to claim 3, characterized in that: The step of training the XGBoost model according to the second training set, obtaining the optimal parameter model through K-fold cross validation, obtaining the stroke prognosis prediction model, and verifying the stroke prognosis prediction model by combining homologous data and non-homologous data according to the first validation set and the second validation set to obtain the prediction model verification result comprises: According to the second training set and the inner-outer K-fold cross validation method, the XGBoost model is trained, and the optimal parameters of the XGBoost model are obtained by grid search to obtain a stroke prognosis prediction model corresponding to the optimal parameters; the optimal parameters include a learning rate and a maximum depth; Performing homologous data verification on the stroke prognosis prediction model according to the first verification set to obtain a corresponding model homologous data verification result; The stroke prognosis prediction model is verified using non-homologous data according to the second verification set to obtain corresponding model heterogeneous data verification results.

7. The method for predicting stroke prognosis according to claim 1, characterized in that: The method further comprises: The SHAP attribution analysis method is used to obtain the feature importance ranking corresponding to the prognosis prediction result, and the corresponding rehabilitation exercise guidance suggestions are generated according to the feature importance ranking.

8. A stroke prognosis prediction system, characterized in that: The system comprises: A data acquisition module is used to acquire characteristic data to be analyzed of stroke patients; the characteristic data to be analyzed includes age, gender, place of origin, mRS score at discharge, stroke-related medical history characteristics, and NIHSS score change characteristics at admission and discharge; The prognosis prediction module is used to input the data to be analyzed into a pre-built stroke prognosis prediction model for prognosis prediction to obtain a corresponding prognosis prediction result; the stroke prognosis prediction model is trained based on a data set with the same characteristics as the characteristic data to be analyzed.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images