PGT-A outcome prediction method for Gn dose recommendation

Abstract

The invention discloses a PGT-A outcome prediction method for Gn dose recommendation, and belongs to the technical field of prediction, and the method comprises the steps: obtaining the physiological indexes of a patient, carrying out the preprocessing of the physiological indexes of the patient, and generating a preprocessing data set; constructing two gradient lifting regression models and a random forest classification model; training the two gradient boosting regression models and one random forest classification model through the preprocessed data set to obtain an ovum total number prediction model, an euploid blastocyst number prediction model and a severe OHSS probability prediction model; and generating a prediction result based on the ovum total number prediction model, the euploid blastocyst number prediction model and the severe OHSS probability prediction model, and feeding back the prediction result to a doctor. Early warning can be provided in the early stage of treatment, the risk of OHSS occurrence is reduced, and the safety of a patient is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the field of prediction technology, and in particular relates to a PGT-A outcome prediction method for Gn dosage recommendation. Background Art

[0002] As an auxiliary selection tool, preimplantation aneuploidy genetic testing (PGT-A) has been shown to significantly increase the live birth rate of each embryo transfer while minimizing the risk of clinical miscarriage, persistent aneuploid pregnancy and implantation failure. The PGT-A process includes the following steps: first, the ovaries are stimulated by drugs to produce and obtain multiple eggs, followed by in vitro fertilization (IVF). Subsequently, after the embryos develop to a certain cell division stage, a small number of embryonic cells are taken for genomic analysis to determine the normality of their chromosome number and structure. Through this screening, embryos carrying chromosomal abnormalities can be screened out and euploid blastocysts that can be transplanted can be obtained, thereby improving the success rate of transplantation and pregnancy rate. For patients with high-risk factors, such as advanced female age, recurrent miscarriages, and recurrent implantation failures, PGT-A may be a more effective treatment option. The success rate of embryo transfer mainly depends on the quality and quantity of transplantable embryos, which is directly affected by individual patient characteristics and ovulation induction treatment regimens. On the one hand, collecting more oocytes can increase the chance of obtaining transplantable euploid embryos, which usually requires the use of higher doses of initial gonadotropin (Gn) to achieve this. On the other hand, the use of high-dose Gn may lead to more intense ovarian stimulation, which increases the cost of treatment and the risk of iatrogenic complications such as moderate to severe ovarian hyperstimulation syndrome (OHSS). In addition, patients undergoing PGT-A usually have more complex and diverse characteristics, which increases the uncertainty of treatment outcomes. Therefore, the development of an ovarian stimulation plan that focuses on an individualized, cost-effective initial dose of Gn is of great clinical and economic significance.

[0003] However, the current methods all assume that the patient's response to the Gn dose is linear. However, in reality, the patient's physiological response may be more complex and may be affected by many factors, and its curve is nonlinear. Therefore, for some patients, the linear model may not accurately predict their actual egg maturation. In addition, ovarian stimulation therapy may be affected by many factors in actual application, such as treatment cycle, ovulation promotion scheme, height and weight, etc., which were not included in the data set in this study. In addition, this study did not study the interpretability of the model, which further limited the research. In addition, in fact, depending on the patient's constitution, the curve should be a linear or convex function, and there is a dose that can obtain the maximum number of eggs. This method does not explicitly model the possibility of avoiding the side effect OHSS. Although doctors can limit the number of eggs obtained through experience to minimize its possibility, due to differences between individuals, experience may not be accurate. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a PGT-A outcome prediction method for Gn dosage recommendation to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides a PGT-A outcome prediction method for Gn dosage recommendation, comprising:

[0006] Acquire physiological indicators of the patient, preprocess the physiological indicators of the patient, and generate a preprocessed data set;

[0007] Build two gradient boosting regression models and one random forest classification model;

[0008] The two gradient boosting regression models and one random forest classification model are trained by using the preprocessed data set to obtain a total egg number prediction model, a euploid blastocyst number prediction model and a severe OHSS probability prediction model;

[0009] Based on the total egg number prediction model, euploid blastocyst number prediction model and severe OHSS probability prediction model, prediction results are generated and fed back to the doctor.

[0010] Preferably, the patient's physiological indicators include: female age, number of follicles, basal gonadotropin level, anti-fluminoic acid level, ovulation induction regimen, height, weight, BMI, initial Gn dose and treatment results.

[0011] Preferably, the process of preprocessing the patient's physiological indicators includes: filling the missing values ​​and abnormal values ​​of the patient's physiological indicators by median to generate the preprocessed data set.

[0012] Preferably, the process of obtaining a total egg number prediction model, a euploid blastocyst number prediction model and a severe OHSS probability prediction model comprises:

[0013] Dividing the preprocessed data set into a training set and a test set based on a five-fold cross validation method;

[0014] The two gradient boosting regression models and one random forest classification model are trained based on the training set, and the models are verified by the test set to generate the total egg number prediction model, euploid blastocyst number prediction model and severe OHSS probability prediction model.

[0015] Preferably, after verifying the model through the test set, the method further includes: gradually removing the input features of the verified model, evaluating the impact of each feature on the model performance, selecting the input feature combination with the best performance, and retraining and evaluating the model.

[0016] Preferably, after verifying the model through the test set, the method further includes: using absolute error and mean square error to evaluate the trained gradient boosting regression model;

[0017] The accuracy rate was used to evaluate the random forest classification model.

[0018] Preferably, the random forest classification model uses a Gini importance calculation method to evaluate the contribution of features to model prediction.

[0019] Preferably, the process of generating prediction results based on the total egg number prediction model, the euploid blastocyst number prediction model and the severe OHSS probability prediction model and feeding back to the doctor includes:

[0020] Predicting the total number of eggs of the patient based on the total egg number prediction model, predicting the number of euploid blastocysts of the patient based on the euploid blastocyst number prediction model, and predicting the probability of OHSS based on the severe OHSS probability prediction model;

[0021] The total number of the patient's eggs, the number of euploid blastocysts and the probability of occurrence of OHSS are fed back to the doctor.

[0022] Compared with the prior art, the present invention has the following advantages and technical effects:

[0023] Traditional linear models assume that the patient's response to Gn dose is linear, but in fact, the patient's physiological response may be nonlinear. The present invention can capture this nonlinear relationship, improve the prediction accuracy of the number of eggs obtained, and help formulate a more accurate ovulation induction plan.

[0024] In the PGT-A process, in order to ensure the success rate, the number of euploid blastocysts that can be transplanted is the key. The present invention can accurately predict the number of euploid blastocysts according to the patient's physiological characteristics and Gn dose, thereby helping doctors to better select suitable embryos for transplantation, and giving patients a psychological expectation of cost and success rate, and improving the success rate of PGT-A.

[0025] By predicting the probability of OHSS occurrence, the present invention can provide early warning in the early stage of treatment, help doctors adjust the Gn dosage, reduce the risk of OHSS occurrence, and ensure the safety of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0027] Figure 1 This is a flow chart of the training and prediction of the initial Gn dose prediction model according to an embodiment of the present invention;

[0028] Figure 2 It is a partial dependence diagram for predicting the total number of eggs according to an embodiment of the present invention;

[0029] Figure 3 This is a partial dependence diagram for predicting the number of euploid blastocysts according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0031] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0032] Embodiment 1

[0033] like Figure 1 As shown, this embodiment provides a PGT-A outcome prediction method for Gn dose recommendation, comprising:

[0034] Acquire physiological indicators of the patient, preprocess the physiological indicators of the patient, and generate a preprocessed data set;

[0035] Build two gradient boosting regression models and one random forest classification model;

[0036] The two gradient boosting regression models and one random forest classification model are trained by using the preprocessed data set to obtain a total egg number prediction model, a euploid blastocyst number prediction model and a severe OHSS probability prediction model;

[0037] Based on the total egg number prediction model, euploid blastocyst number prediction model and severe OHSS probability prediction model, prediction results are generated and fed back to the doctor.

[0038] The specific process includes:

[0039] (1) Data collection and preprocessing: Doctors collected data including the patient's physiological indicators such as the woman's age, number of follicles (AFC), basal gonadotropin level (bFSH), anti-fluoroquinolone level (AMH), ovulation induction regimen, height, weight, body mass index (BMI), initial Gn dose, and treatment results (number of oocytes obtained, number of euploid blastocysts, and occurrence of OHSS). Data cleaning was then performed, and the median was used to fill in missing values ​​and outliers.

[0040] (2) Model training: Split the dataset into 80% training set and 20% test set, use five-fold cross validation, and select the best performing dataset split. Input the data of the training set, including the patient's physiological indicators and initial Gn dose. Train a gradient boosting regression model to predict the total number of eggs, and the model outputs the total number of eggs. Train a gradient boosting model to predict the number of euploid blastocysts, and the model outputs the number of euploid blastocysts. Train a random forest classification model to predict the probability of moderate to severe OHSS, and the model outputs the probability of moderate to severe OHSS.

[0041] (3) Model evaluation: The performance of the three models is evaluated using the test set data. For the regression model, the evaluation indicators MAE and RMSE are calculated. For the classification model, the accuracy (ACC) is calculated.

[0042] (4) Ablation experiment: Conduct an ablation experiment to evaluate the impact of each feature on model performance by gradually removing input features. Select the input feature combination with the best performance, retrain and evaluate the model.

[0043] (5) Interpretability evaluation: First, we perform feature importance analysis to identify the features that have the greatest impact on the prediction results. Next, we use partial dependency plots to further analyze these key features to understand the specific impact of different values ​​on the prediction results.

[0044] (6) Prediction: For new patients, input their physiological indicators and traverse the initial Gn dose. Use the trained gradient boosting regression model to predict the number of oocytes retrieved and the number of euploid blastocysts. Use the trained random forest classification model to predict the probability of OHSS. Feedback the prediction results to the doctor to assist in formulating a personalized treatment plan.

[0045] This embodiment proposes three machine learning models. First, the first gradient regression boosting model solves the problem of predicting the total number of eggs at a specified starting Gn dose. The second gradient regression boosting model solves the key to the problem, predicting the number of transplantable embryos based on the starting Gn dose. In addition, this embodiment establishes a random forest classification model for predicting the probability of OHSS. Based on the output of these three models, doctors can make better decisions and give patients a complete expectation.

[0046] First is the first gradient regression boosting model. The gradient boosting regression model optimizes model performance by iteratively constructing multiple decision trees, and can obtain high prediction accuracy in a small sample size data set. It is good at capturing complex nonlinear relationships in data and has a certain robustness to data noise. In addition, the model also has feature selection capabilities and can identify the variables that are most important to the prediction results. The first gradient regression boosting model proposed in this embodiment is different from other models in predicting the slope of the linear dose-response function. The output is the total number of eggs. This method can better fit the drug response function because in fact the function is often non-monotonic. There may be two different solutions for the starting Gn dosage to obtain the specified total number of eggs. These two solutions may exist on both sides of the extreme point, either low dose or high dose. At the same time, high dose also brings the risk of side effects. By fitting and predicting the response curve, the optimal solution can be given according to the actual situation. The first gradient regression boosting model simulates the ovulation induction process of PGT. The second is the second gradient regression boosting model, which also uses the gradient boosting model, but unlike the first gradient regression boosting model, the model outputs the number of transplantable embryos.

[0047] Based on actual needs, in order to ensure the success rate of PGT-A, a starting dose that can produce one or more euploid blastocysts is often selected, which means that more eggs will be produced, but the probability of side effects will also increase. Therefore, a random forest classification model was established. Unlike the previous two models, this model uses a random forest classification model. Similar to the gradient boosting model, the random forest classifier can effectively capture the nonlinear relationship in the data by constructing multiple decision trees and combining their prediction results, and it is interpretable. The random forest classification model inputs physiological indicators and the starting Gn dose to predict the probability of OHSS at this dose.

[0048] In this study, this embodiment uses the Gini importance calculation method of the model to evaluate the contribution of features to model prediction. Since the models are all tree models, the feature screening process based on Gini importance is determined by analyzing the Gini impurity reduced when each feature splits the node in the model. According to the evaluation results of Gini importance and experimental results, this embodiment selects some features with less importance for deletion, thereby optimizing the prediction ability and generalization performance of the model. And by observing the importance, the model has a certain degree of interpretability.

[0049] The dataset contains 2506 data from 2014 to 2024, including 1852 collected from the First Affiliated Hospital of Sun Yat-sen University and 654 collected from Nanning People's Hospital.

[0050] The data set consists of a variety of related factors, mediating variables, PGT indications, outcome variables, and PGT indication-related characteristics. The related factors include female age, number of follicles (AFC), basal gonadotropin level (bFSH), anti-fluoroquinolone level (AMH), ovulation induction regimen, height, weight, BMI, basal luteinizing hormone level (bLH), estradiol level (E2), FSH / LH, male age, number of spontaneous abortions, and previous live births. The mediating variable is the initial Gn dose, and the outcome variable includes the total number of eggs, the number of transplantable embryos, and whether it is severe OHSS. Finally, the PGT indication-related characteristics include whether it is severe oligoasthenozoospermia, whether it is PCOS, and PGT indications. Among them, ovulation induction regimen, whether it is severe OHSS, whether it is severe oligoasthenozoospermia, whether it is PCOS, and PGT indications, and the rest are continuous variables. In addition, BMI and FSH / LH are calculated based on basic characteristics.

[0051] In addition, this embodiment compares the proposed first gradient regression boosting model with the model for predicting the slope of the response function, both of which use the same feature input. And given that the total number of eggs and the number of euploid blastocysts are positively correlated when other features are the same, this embodiment uses the slope model to take the number of transplantable embryos as the fitting target and compares it with the performance of the second gradient regression boosting model. This embodiment randomly selects five samples, obtains the function predicted by the model by adjusting the important features and Gn dosage, and examines the fitting and learning effects of the model.

[0052] Table 1 shows the performance of the model on the test set. According to the experimental results, it can be observed that Model 1 achieved 4.19 in MAE and 5.67 in RMSE, which is better than the Pova model's 4.43 and 6.27 and Slope's 4.48 and 6.35. Model 2 achieved 1.19 and 1.71 in MAE and RMSE respectively. Finally, the classification model Model 3 achieved an ACC of 0.99.

[0053] Table 1

[0054]

[0055]

[0056] In order to make the model more interpretable, Table 2 reports the importance of the model input features and the features selected for prediction in this embodiment, and the important features are marked. Not all features in the dataset are selected as input. This embodiment conducts ablation experiments based on importance. The selected features can achieve the best effect, increasing or decreasing the lossy model performance.

[0057] Table 2

[0058]

[0059] In order to examine whether the model can learn the important features and the relationship between the initial Gn dose and the output, this example draws a partial dependency graph of the first gradient regression boosting model, such as Figure 2 As shown, not all patients have linear response curves. There may be an extreme point where the most eggs can be obtained when this dose is used. Increasing or decreasing the dose will reduce the number. This also confirms that for the recommended starting Gn dose, a nonlinear model should be used. In addition, it can be observed that some patients are insensitive to the drug, and increasing or decreasing the dose has little effect on the output.

[0060] In addition, if Figure 3 As shown, this embodiment also draws a partial dependence graph of the second gradient regression boosting model. It can be observed that the trend of the curve is similar to the partial dependence graph of the first gradient regression boosting model.

[0061] The above content proves that the method of this embodiment is superior to the existing method, and can capture the nonlinear characteristics of the drug response curve, and can effectively predict the number of euploid blastocysts in PGT-A.

[0062] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A PGT-A outcome prediction method for Gn dosage recommendation, characterized in that: The following steps are involved: Acquire physiological indicators of the patient, preprocess the physiological indicators of the patient, and generate a preprocessed data set; Build two gradient boosting regression models and one random forest classification model; The two gradient boosting regression models and one random forest classification model are trained by using the preprocessed data set to obtain a total egg number prediction model, a euploid blastocyst number prediction model and a severe OHSS probability prediction model; Based on the total egg number prediction model, euploid blastocyst number prediction model and severe OHSS probability prediction model, prediction results are generated and fed back to the doctor.

2. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 1, characterized in that: The patient's physiological indicators include: female age, number of follicles, basal gonadotropin level, anti-fluoroquinolone level, ovulation induction regimen, height, weight, BMI, initial Gn dose and treatment results.

3. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 1, characterized in that: The process of preprocessing the patient's physiological indicators includes: filling the missing values ​​and abnormal values ​​of the patient's physiological indicators with the median to generate the preprocessed data set.

4. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 1, characterized in that: The process of obtaining the total egg number prediction model, the euploid blastocyst number prediction model and the severe OHSS probability prediction model comprises: Dividing the preprocessed data set into a training set and a test set based on a five-fold cross validation method; The two gradient boosting regression models and one random forest classification model are trained based on the training set, and the models are verified by the test set to generate the total egg number prediction model, euploid blastocyst number prediction model and severe OHSS probability prediction model.

5. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 4, characterized in that: After verifying the model through the test set, the method also includes: gradually removing the input features of the verified model, evaluating the impact of each feature on the model performance, selecting the input feature combination with the best performance, and retraining and evaluating the model.

6. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 4, characterized in that: After verifying the model through the test set, the method further includes: using absolute error and mean square error to evaluate the trained gradient boosting regression model; The accuracy rate was used to evaluate the random forest classification model.

7. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 4, characterized in that: The random forest classification model uses the Gini importance calculation method to evaluate the contribution of features to model prediction.

8. The PGT-A outcome prediction method for Gn dosage recommendation according to claim 1, characterized in that: The process of generating prediction results based on the total egg number prediction model, the euploid blastocyst number prediction model and the severe OHSS probability prediction model and feeding back to the doctor includes: Predicting the total number of eggs of the patient based on the total egg number prediction model, predicting the number of euploid blastocysts of the patient based on the euploid blastocyst number prediction model, and predicting the probability of OHSS based on the severe OHSS probability prediction model; The total number of the patient's eggs, the number of euploid blastocysts and the probability of occurrence of OHSS are fed back to the doctor.

Images