Fast peritoneal solute transport rate probability prediction method, apparatus, device, and medium

Abstract

The application discloses a rapid peritoneal solute transport rate probability prediction method, device, equipment and medium. The method comprises the following steps: acquiring target clinical feature data of a target user; and inputting the target clinical feature data into a trained target machine learning model to predict a rapid peritoneal solute transport rate probability of the target user. The technical scheme provided by the application directly predicts the rapid peritoneal solute transport rate probability by using the pre-trained machine learning model, realizes more convenient and efficient PSTR evaluation, and thus makes up for the operation limitation of PET, helps to early identify rapid PSTR, and helps some primary medical and health units which are not available for PET to preliminarily evaluate the peritoneal transport rate of patients.

Description

Technical Field

[0001] This invention relates to the field of peritoneal dialysis technology, and in particular to a method, apparatus, equipment and medium for probabilistic prediction of rapid peritoneal solute transport rate. Background Technology

[0002] Chronic kidney disease has become a significant burden on global public health. Currently, approximately 3.8 million people worldwide rely on dialysis to treat end-stage renal disease, with about 11% of these patients opting for peritoneal dialysis (PD). The use of peritoneal dialysis is rapidly increasing in many regions globally, showing a year-on-year upward trend. Compared to hemodialysis, peritoneal dialysis offers greater treatment autonomy and a lower treatment burden, while also possessing advantages in maintaining hemodynamic stability and protecting residual renal function. Although PD is a home-based treatment, regular hospital visits to assess peritoneal function are crucial for treatment success. Timely identification of the peritoneal solute transport rate (PSTR) helps optimize and individualize PD regimens.

[0003] Currently, the peritoneal equilibrium test (PET) is commonly used to assess peritoneal dialysis (PSTR). Besides the classic PET, several other PET testing methods have emerged in recent years, such as modified PET, miniature PET, and dual-miniature PET. However, these methods all require complex procedures, demand high levels of professional expertise from nursing staff, and require patients to stay in the hospital for more than 4 hours for follow-up, increasing follow-up time and medical workload, thus limiting their widespread adoption in primary hospitals (especially in developing countries). Furthermore, although machine learning models have been applied in peritoneal dialysis, they typically focus on adverse outcomes, and no research has yet addressed the prediction of peritoneal functional status. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and medium for predicting the probability of rapid peritoneal solute transport rate, in order to overcome the operational limitations of PET and help identify rapid PSTRs at an early stage.

[0005] In a first aspect, embodiments of the present invention provide a method for probabilistic prediction of rapid peritoneal solute transport rates, the method comprising: Obtain target clinical characteristic data of the target users; The target clinical feature data is input into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

[0006] Optionally, before inputting the target clinical feature data into the trained target machine learning model to predict the probability of the target user's rapid peritoneal solute transport rate, the method further includes: Obtain peritoneal equilibration test results and corresponding clinical characteristic data of the samples, and divide them into training set and validation set; The target machine learning model is trained using the training set, and the trained target machine learning model is tested using the validation set to obtain target model performance data. The number of feature types in the sample clinical feature data is gradually reduced, and the training and testing process is repeated to determine the feature types to be selected for the target clinical feature data based on the changing trend of the target model performance data.

[0007] Optionally, after determining the feature type selected for the target clinical feature data based on the changing trend of the target model performance data, the method further includes: The SHAP interpretation method was used to determine the contribution of each type of feature in the target clinical feature data to the prediction of rapid peritoneal solute transport rate, and the influence of the value of each type of feature in the target clinical feature data on whether it is a rapid peritoneal solute transport rate.

[0008] Optionally, the target clinical characteristic data may include the following feature types: 24-hour dialysate / plasma creatinine ratio, overnight post-abdominal dialysate / plasma creatinine ratio, overnight ultrafiltration volume, serum total protein, and peritoneal dialysis age.

[0009] Optionally, after obtaining the peritoneal equilibration test results data and corresponding sample clinical characteristic data, and dividing them into training and validation sets, the method further includes: Multiple candidate machine learning models are trained based on the training set, and the trained candidate machine learning models are tested based on the validation set to obtain the target model performance data for each candidate machine learning model. The target model performance data of each of the candidate machine learning models are compared to determine the target machine learning model to be used.

[0010] Optionally, the target machine learning model is the XGBoost model.

[0011] Secondly, embodiments of the present invention also provide a rapid peritoneal solute transport rate probability prediction device, the device comprising: The target data acquisition module is used to acquire the target clinical characteristic data of the target user; The probability prediction module is used to input the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

[0012] Thirdly, embodiments of the present invention also provide a computer device, the computer device comprising: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention.

[0013] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention.

[0014] Fifthly, embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention.

[0015] This invention provides a method for predicting the probability of rapid peritoneal solute transport rate (PSTR). First, target clinical feature data of the target user is acquired. Then, the obtained target clinical feature data is input into a trained target machine learning model to predict the probability of PSTR for the target user. This method, by directly predicting the probability of PSTR using a pre-trained machine learning model, achieves a simpler and more efficient PSTR assessment, thus overcoming the operational limitations of PET. It can help identify rapid PSTR early and assist primary healthcare units in areas where PET is inaccessible in initially assessing the peritoneal transport rate of patients. Attached Figure Description

[0016] Figure 1 A flowchart of the rapid peritoneal solute transport rate probability prediction method provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram illustrating the contribution of each feature provided in Embodiment 1 of the present invention to the final model; Figure 3 This is a SHAP bee colony diagram provided in Embodiment 1 of the present invention; Figure 4 This is the SHAP dependency graph provided in Embodiment 1 of the present invention; Figure 5 This is a schematic diagram illustrating the performance of nine machine learning models provided in Embodiment 1 of the present invention in recognizing fast PSTRs. Figure 6This is a schematic diagram of the rapid peritoneal solute transport rate probability prediction device provided in Embodiment 2 of the present invention; Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment 3 of the present invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0018] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the steps as sequential processes, many of these steps can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the steps can be rearranged. The process can be terminated when its operation is complete, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0019] Example 1 Figure 1 This is a flowchart of a rapid peritoneal solute transport rate probability prediction method provided in Embodiment 1 of the present invention. This embodiment is applicable to situations requiring assessment of the peritoneal solute transport rate in peritoneal dialysis patients. This method can be executed by the rapid peritoneal solute transport rate probability prediction device provided in this embodiment of the invention. This device can be implemented in hardware and / or software, and is generally integrated into a computer device. Figure 1 As shown, the method specifically includes the following steps: S11. Obtain target clinical characteristic data of the target user.

[0020] S12. Input the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

[0021] Specifically, the target user can be any peritoneal dialysis patient requiring follow-up assessment of peritoneal solute transport rate (PSTR). Target clinical characteristic data can be selected from their relevant baseline data for prediction. Preferably, the target clinical characteristic data includes 24-hour dialysate / plasma creatinine ratio, overnight dialysis fluid / plasma creatinine ratio, overnight ultrafiltration volume, serum total protein, and peritoneal dialysis age. After obtaining the target clinical characteristic data, it can be input into a trained target machine learning model to automatically predict the probability of the target user's rapid peritoneal solute transport rate. Preferably, the target machine learning model is the XGBoost model. Furthermore, the PSTR status of the target user can be assessed based on the obtained rapid peritoneal solute transport rate probability. For example, if the predicted rapid peritoneal solute transport rate probability is 54.84%, it can be further assessed that the current target user is more likely to have a rapid peritoneal solute transport rate status; if the predicted rapid peritoneal solute transport rate probability is 42.63%, it can be further assessed that the current target user is more likely to have a non-rapid peritoneal solute transport rate status.

[0022] To promote the practical application of the model in clinical settings, the trained target machine learning model can be deployed as an interactive web application using the Shiny package in R. The application interface can include input boxes for various target clinical feature data and prediction buttons. Users can input the obtained target clinical feature data and click the prediction button. The program will then automatically generate a rapid PSTR probability and display it on the application interface. Furthermore, it can assess the PSTR status based on the obtained rapid PSTR probability and display that assessment result online as well. Users can then easily and quickly obtain the PSTR status assessment result online, allowing them to refer to the assessment result for further examination or treatment.

[0023] Based on the above technical solution, optionally, before inputting the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user, the method further includes: acquiring peritoneal equilibration test result data and corresponding sample clinical feature data, and dividing them into training set and validation set; training the target machine learning model according to the training set, and testing the trained target machine learning model according to the validation set to obtain target model performance data; gradually reducing the feature types of the sample clinical feature data, and repeating the training and testing process, so as to determine the feature types selected for the target clinical feature data according to the changing trend of the target model performance data.

[0024] Specifically, the required dataset can be constructed based on historically completed PET data. For example, laboratory data and dialysis adequacy assessment data (Kt / V) from PET examinations performed at a hospital between 2006 and 2023 could be collected. The hospital's follow-up plan could be: patients routinely undergo a standard PET examination every 6 months, and dialysis adequacy assessment and comprehensive laboratory tests are performed every 3 months during the follow-up period. Furthermore, cases with incomplete PET data and missing Kt / V data can be excluded from the collected data to obtain standard peritoneal equilibration test results for constructing a predictive model. Then, corresponding clinical characteristic data of the peritoneal equilibration test results can be collected. These clinical characteristic data can include demographic characteristics and clinical features, etc. For example, 53 variables are included as follows: age, sex, body surface area, peritoneal dialysis age, overnight stay time, overnight ultrafiltration volume, 24-hour ultrafiltration volume, urine volume, serum creatinine (SCr), overnight dialysis fluid creatinine, 24-hour dialysis fluid creatinine, urine creatinine, nighttime dialysis fluid / plasma creatinine ratio (night D / Pcr), 24-hour dialysis fluid / plasma creatinine ratio (day D / Pcr), BUN, nighttime dialysis fluid urea nitrogen, 24-hour dialysis fluid urea nitrogen, urine BUN, 24-hour dialysate-to-plasma urea concentration ratio (day D / Pur), nighttime dialysis fluid / plasma urea nitrogen ratio (night D / Pur), GFR, and total creatinine clearance. (Chronic acid rate, CCr), residual kidney CCr, total Kt / v, residual kidney Kt / v, total protein (TP), Alb, serum globulin, serum glucose, serum calcium, serum phosphorus, serum sodium, serum chloride, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, uric acid, serum carbon dioxide, total cholesterol, triglycerides, low-density lipoprotein, high-density lipoprotein, Hb, hematocrit, red blood cells, white blood cells, platelets, neutrophils, granulocytes, lymphocytes, monocytes, eosinophils, and basophils.

[0025] Furthermore, clinical characteristic data of the samples can be used as model input, and peritoneal balancing test results data as model output. The large amount of collected data can be divided into training and validation sets, specifically in a ratio of 75% and 25%. For example, 5617 cases of raw PET data were collected, and 581 cases with incomplete PET data and missing corresponding Kt / V data were excluded, resulting in 5036 cases of peritoneal balancing test results data that met the standards. After obtaining the corresponding clinical characteristic data of the samples, 3777 pairs of training data and 1259 pairs of validation data were obtained.

[0026] The target machine learning model can then be trained based on the obtained training set, and tested based on the obtained validation set to obtain target model performance data. The target model performance data may include at least one of the following: Area Under the Receiver Operating Characteristic Curve (AUROC), accuracy, specificity, sensitivity, positive predictive value, negative predictive value, and F1 index.

[0027] Feature selection can then be performed. Initially, all 53 variables mentioned above can be used for model building. The modeling process can employ ten-fold cross-validation to enhance the model's generalization ability and robustness. After completing the initial training, the feature types of the sample clinical feature data can be gradually reduced according to feature importance, such as removing some variables from all 53 variables. The training and testing process can then be repeated based on the reduced sample clinical feature data. Simultaneously, the performance data of the target model obtained from the tests can be observed. When the performance data of the target model shows a significant decline, the currently used feature types can be used as the feature types for the target clinical feature data. At the same time, the currently trained model can be used as the trained target machine learning model, or further training can be performed. Taking the XGBoost model as an example of a target machine learning model, tests showed that when the number of features was reduced to 5, the model's AUROC, specificity, sensitivity, and F1 index all decreased significantly. The final 5 features included in the model were 24-hour dialysate / plasma creatinine ratio (day D / Pcr), nighttime dialysis fluid / plasma creatinine ratio (night D / Pcr), nighttime ultrafiltration volume (Night UF), serum total protein (TP), and peritoneal dialysis age (PD Vintage). At this point, the model achieved an AUROC of 0.871, accuracy of 0.795, specificity of 0.762, sensitivity of 0.817, positive predictive value of 0.836, negative predictive value of 0.738, and F1 index of 0.826 when predicting rapid PSTRs.

[0028] Optionally, after determining the feature type selected for the target clinical feature data based on the changing trend of the target model performance data, the method further includes: using the SHAP interpretation method to determine the contribution of each type of feature in the target clinical feature data to the prediction of rapid peritoneal solute transport rate and the influence relationship between the value of each type of feature in the target clinical feature data and whether it is a rapid peritoneal solute transport rate.

[0029] Specifically, combining machine learning with the SHAP (SHapley Additive exPlanation) interpretation method can improve the interpretability of the constructed target machine learning model, thus facilitating clinical translation. The SHAP interpretation method can provide a consistent and quantifiable contribution assessment for each feature and intuitively present the relationship between input features and fast PSTRs. Based on the XGBoost model and the five selected features—day D / Pcr, night D / Pcr, Night UF, TP, and PD Vintage—the contribution of each feature to the final model is as follows: Figure 2 As shown, the bee colony is sorted from high to low based on the average SHAP value. The SHAP bee colony diagram is as follows. Figure 3 As shown in the figure, this diagram visually presents the positive and negative impacts of each feature value on the prediction result and their distribution. A SHAP dependency graph can be further drawn as follows: Figure 4 As shown, this illustrates the specific relationship between each feature and the prediction result. Figure 4 In the model, SHAP > 0 indicates that this feature has a positive impact on the model's prediction of rapid PSTR, suggesting that such patients have a higher risk of developing rapid PSTR. Conversely, SHAP < 0 indicates that this feature has a negative impact on the model's prediction of rapid PSTR, suggesting that such patients have a lower risk of developing rapid PSTR. Figure 4 The results showed that patients with day D / Pcr>0.8, night D / Pcr>1.00, night UF<300mL, TP<58g / L, or PD Vintage>98 months all had positive SAP values, and patients with these characteristics were more likely to be predicted as rapid PSTR.

[0030] Based on the above technical solution, optionally, after obtaining the peritoneal equilibration test result data and the corresponding sample clinical characteristic data, and dividing them into training set and validation set, the method further includes: training multiple candidate machine learning models according to the training set, and testing the trained candidate machine learning models according to the validation set to obtain the target model performance data corresponding to each candidate machine learning model; comparing the target model performance data of each candidate machine learning model to determine the selected target machine learning model.

[0031] Specifically, to select the optimal prediction model, several candidate machine learning models can be initially developed and tested for comparison. For example, nine candidate machine learning models are included: decision tree (DT), k-nearest neighbors (KNN), lightweight gradient boosting machine (LightGBM), artificial neural network (ANN), random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost), least absolute shrinkage and selection operator (LASSO) regression, and logistic regression. Specifically, while training the target machine learning model, all candidate machine learning models can be trained and tested simultaneously using the same training and validation sets. The target machine learning model is one of the candidate machine learning models. The training process can also include multiple training and testing processes with gradually decreasing feature types, and the performance data results of the target model for each model can be recorded separately. The performance data of the target model obtained from one training and testing session is as follows: Figure 5 As shown, the XGBoost model exhibits the best discriminative ability (AUROC=0.881), followed by the RF model (AUROC=0.878) and the LightGBM model (AUROC=0.878). Although all three models can effectively identify fast PSTRs, after feature selection, the XGBoost model outperforms the other models in overall performance. In contrast, the predictive ability of the RF model and the LightGBM model decreases significantly when the number of features decreases. Therefore, the XGBoost model is the optimal model for use as the target machine learning model in actual prediction.

[0032] Studies have shown that patients undergoing rapid peritoneal dialysis (PSTR) have a higher risk of adverse clinical outcomes (such as mortality and inadequate ultrafiltration) compared to those undergoing non-rapid PSTR. To evaluate the predictive ability of a target-based machine learning model for the clinical prognosis of patients with peripheral dialysis (PD), 2013 patients who had previously started PD treatment and had baseline PET data were selected. Based on the aforementioned prediction method, these patients were divided into rapid PSTR and non-rapid PSTR groups. Simultaneously, a corrected Cox proportional hazards model was used for analysis. The results showed that patients predicted to undergo rapid PSTR had a significantly higher risk of all-cause mortality than those predicted to undergo non-rapid PSTR (HR=1.31, 95% CI 1.04–1.66, P=0.020), and the risk of surgical failure was also significantly increased in patients undergoing rapid PSTR (HR=1.28, 95% CI 1.01–1.62, P=0.046). Therefore, the aforementioned target-based machine learning model has good predictive ability for the clinical outcomes of peritoneal dialysis patients.

[0033] The technical solution provided in this invention first acquires target clinical feature data of the target user, and then inputs the obtained target clinical feature data into a trained target machine learning model to predict the probability of rapid peritoneal solute transport rate (PSTR) for the target user. By directly predicting the probability of PSTR using a pre-trained machine learning model, a simpler and more efficient PSTR assessment is achieved, thereby compensating for the operational limitations of PET and helping to identify rapid PSTR at an early stage. This can assist primary healthcare units that are not accessible by PET in making a preliminary assessment of the peritoneal transport rate of patients.

[0034] Example 2 Figure 6 This is a schematic diagram of the rapid peritoneal solute transport rate probability prediction device provided in Embodiment 2 of the present invention. This device can be implemented in hardware and / or software, and is generally integrated into a computer device to execute the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention. Figure 6 As shown, the device includes: Target data acquisition module 21 is used to acquire target clinical characteristic data of target users; The probability prediction module 22 is used to input the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

[0035] The technical solution provided in this invention first acquires target clinical feature data of the target user, and then inputs the obtained target clinical feature data into a trained target machine learning model to predict the probability of rapid peritoneal solute transport rate (PSTR) for the target user. By directly predicting the probability of PSTR using a pre-trained machine learning model, a simpler and more efficient PSTR assessment is achieved, thereby compensating for the operational limitations of PET and helping to identify rapid PSTR at an early stage. This can assist primary healthcare units that are not accessible by PET in making a preliminary assessment of the peritoneal transport rate of patients.

[0036] Based on the above technical solution, optionally, the device further includes: The sample data acquisition module is used to acquire peritoneal equilibrium test result data and corresponding sample clinical feature data before inputting the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user, and to divide it into training set and validation set. The model training module is used to train the target machine learning model according to the training set and to test the trained target machine learning model according to the validation set to obtain target model performance data. The feature type determination module is used to gradually reduce the feature types of the sample clinical feature data and repeat the training and testing process to determine the feature types selected for the target clinical feature data based on the changing trend of the target model performance data.

[0037] Based on the above technical solution, optionally, the device further includes: The model interpretation module is used to determine the contribution of each type of feature in the target clinical feature data to the prediction of rapid peritoneal solute transport rate, and the influence relationship between the value of each type of feature in the target clinical feature data and whether it is a rapid peritoneal solute transport rate, after the feature type selected for the target clinical feature data is determined based on the trend of the target model performance data.

[0038] Based on the above technical solution, optionally, the target clinical characteristic data may include the 24-hour dialysate / plasma creatinine ratio, the overnight peritoneal dialysis fluid / plasma creatinine ratio, the overnight ultrafiltration volume, serum total protein, and peritoneal dialysis age.

[0039] Based on the above technical solution, optionally, the device further includes: The multi-model training module is used to train multiple candidate machine learning models according to the training set after obtaining peritoneal equilibration test result data and corresponding sample clinical feature data, and dividing them into training set and validation set, and to test the trained candidate machine learning models according to the validation set, so as to obtain the target model performance data of each candidate machine learning model. The target machine learning model selection module is used to compare the target model performance data of each candidate machine learning model in order to determine the selected target machine learning model.

[0040] Optionally, based on the above technical solution, the target machine learning model may be the XGBoost model.

[0041] The rapid peritoneal solute transport rate probability prediction device provided in the embodiments of the present invention can execute the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.

[0042] It is worth noting that in the embodiments of the above-mentioned rapid peritoneal solute transport rate probability prediction device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0043] Example 3 Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment 3 of the present invention, showing a block diagram of an exemplary computer device suitable for implementing the embodiments of the present invention. Figure 7 The computer device shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the present invention. Figure 7 As shown, the computer device includes a processor 31, a memory 32, an input device 33, and an output device 34; the number of processors 31 in the computer device can be one or more. Figure 7 Taking a processor 31 as an example, the processor 31, memory 32, input device 33, and output device 34 in a computer device can be connected via a bus or other means. Figure 7 Taking the example of a connection between China and Israel via a bus.

[0044] The memory 32, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the rapid peritoneal solute transport rate probability prediction method in this embodiment of the invention (e.g., the target data acquisition module 21 and the probability prediction module 22 in the rapid peritoneal solute transport rate probability prediction device). The processor 31 executes various functional applications and data processing of the computer device by running the software programs, instructions, and modules stored in the memory 32, thereby realizing the aforementioned rapid peritoneal solute transport rate probability prediction method.

[0045] The memory 32 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 32 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory, or other non-volatile solid-state storage device. In some instances, the memory 32 may further include memory remotely located relative to the processor 31, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0046] Input device 33 can be used to acquire target clinical characteristic data of the target user, and to generate key signal inputs related to user settings and function control of the computer device. Output device 34 may include a display screen, which can be used to display probability prediction results to the user, etc.

[0047] Example 4 Embodiment 4 of the present invention also provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform a rapid peritoneal solute transport rate probability prediction method, the method comprising: Obtain target clinical characteristic data of the target users; The target clinical feature data is input into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

[0048] Storage media can be any type of memory device or storage device. The term "storage media" is intended to include: mounting media, such as CD-ROMs, floppy disks, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disks or optical storage); registers or other similar types of memory elements. Storage media may also include other types of memory or combinations thereof. Furthermore, storage media may reside in a computer system in which the program is executed, or may reside in a different second computer system connected to the computer system via a network (such as the Internet). The second computer system can provide program instructions to the computer for execution. The term "storage media" can include two or more storage media that may reside in different locations (e.g., in different computer systems connected via a network). Storage media may store program instructions (e.g., specifically implemented as a computer program) that can be executed by one or more processors.

[0049] Of course, the computer-executable instructions provided in the embodiments of the present invention are not limited to the method operations described above, but can also perform related operations in the rapid peritoneal solute transport rate probability prediction method provided in any embodiment of the present invention.

[0050] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0051] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0052] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0053] Example 5 Embodiment 5 of the present invention also provides a computer program product, which includes a computer program (also referred to as code or instructions). The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it is used to execute the rapid peritoneal solute transport rate probability prediction method provided in any of the above embodiments, and has the corresponding beneficial effects of executing the method.

[0054] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for probabilistic prediction of rapid peritoneal solute transport rate, characterized in that, include: Obtain target clinical characteristic data of the target users; The target clinical feature data is input into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

2. The method for probabilistic prediction of rapid peritoneal solute transport rate according to claim 1, characterized in that, Before inputting the target clinical feature data into the trained target machine learning model to predict the probability of the target user's rapid peritoneal solute transport rate, the method further includes: Obtain peritoneal equilibration test results and corresponding clinical characteristic data of the samples, and divide them into training set and validation set; The target machine learning model is trained using the training set, and the trained target machine learning model is tested using the validation set to obtain target model performance data. The number of feature types in the sample clinical feature data is gradually reduced, and the training and testing process is repeated to determine the feature types to be selected for the target clinical feature data based on the changing trend of the target model performance data.

3. The method for probabilistic prediction of rapid peritoneal solute transport rate according to claim 2, characterized in that, After determining the feature type selected for the target clinical feature data based on the changing trend of the target model performance data, the method further includes: The SHAP interpretation method was used to determine the contribution of each type of feature in the target clinical feature data to the prediction of rapid peritoneal solute transport rate, and the influence of the value of each type of feature in the target clinical feature data on whether it is a rapid peritoneal solute transport rate.

4. The method for probabilistic prediction of rapid peritoneal solute transport rate according to claim 1 or 2, characterized in that, The target clinical characteristic data selected include the 24-hour dialysate / plasma creatinine ratio, the overnight peritoneal dialysis fluid / plasma creatinine ratio, the overnight ultrafiltration volume, serum total protein, and peritoneal dialysis age.

5. The method for probabilistic prediction of rapid peritoneal solute transport rate according to claim 2, characterized in that, After obtaining the peritoneal equilibration test results data and corresponding sample clinical characteristic data, and dividing them into training and validation sets, the process further includes: Multiple candidate machine learning models are trained based on the training set, and the trained candidate machine learning models are tested based on the validation set to obtain the target model performance data for each candidate machine learning model. The target model performance data of each of the candidate machine learning models are compared to determine the target machine learning model to be used.

6. The method for probabilistic prediction of rapid peritoneal solute transport rate according to claim 1 or 5, characterized in that, The target machine learning model is the XGBoost model.

7. A device for predicting the probability of rapid peritoneal solute transport rate, characterized in that, include: The target data acquisition module is used to acquire the target clinical characteristic data of the target user; The probability prediction module is used to input the target clinical feature data into the trained target machine learning model to predict the probability of the rapid peritoneal solute transport rate of the target user.

8. A computer device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the rapid peritoneal solute transport rate probability prediction method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the rapid peritoneal solute transport rate probability prediction method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the rapid peritoneal solute transport rate probability prediction method as described in any one of claims 1-6.

Images