Neural development disorder prediction model construction method, system, equipment and medium

Abstract

The invention provides a neural developmental disorder prediction model construction method, system and device and a medium, and the method comprises the steps: obtaining the clinical information of a child patient, the clinical information comprises severity of developmental retardation / dyspepsia, complexity of syndromic neurodevelopmental disorder, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain development deformity, hearing / visual function impairment and related ear-eye development deformity; applying logistic regression analysis to screen out predictive factors significantly related to the positive diagnosis rate of third-generation whole exon genome sequencing in the clinical information; constructing a neural developmental disorder prediction model based on binary logistic regression according to the prediction factor; and verifying the neurodevelopmental disorder prediction model by applying data of the internal verification group and the external verification group so as to evaluate the prediction efficiency of the neurodevelopmental disorder prediction model, so that the probability of obtaining positive diagnosis after the child patient receives trio-WES detection can be accurately evaluated, and the prediction efficiency is remarkably improved.

Description

Technical Field

[0001] This application relates to the technical field of neurodevelopmental disorder syndrome prediction, and in particular, to a method, system, electronic device, and storage medium for constructing a neurodevelopmental disorder prediction model. Background Art

[0002] Syndromic neurodevelopmental disorders (s-NDDs) are a group of neurological developmental diseases with complex etiologies and diverse clinical manifestations. They mainly affect the neurological development of children and usually manifest as disorders in multiple aspects such as cognition, language, and movement. These disorders may be accompanied by other mental or neurological symptoms, such as autism spectrum disorder (ASD), epilepsy (EP), and attention deficit hyperactivity disorder (ADHD). The etiology of s-NDDs has not been fully clarified, and the manifestations of the diseases show a high degree of heterogeneity, making diagnosis and treatment more difficult. Early diagnosis of these diseases is crucial because early intervention has a significant impact on improving the prognosis and quality of life of children. With the development of genomics, especially the application of third-generation genomic sequencing technology (trio-WES), the etiology of some s-NDDs patients has been clarified, but the overall diagnostic rate is still low, making it difficult to meet the clinical demand for early and accurate diagnosis.

[0003] Currently, although trio-WES technology has made significant progress in the diagnosis of some diseases, its application in s-NDDs still faces challenges. On the one hand, due to the high phenotypic diversity of s-NDDs, the symptoms of many children are atypical, which brings difficulties to the determination of positive results in genetic testing. On the other hand, the cost of trio-WES is relatively high, and the detection process usually takes a long time, which limits its popularization in clinical applications to a certain extent. Therefore, there is a lack of a prediction model based on clinical phenotype factors to estimate the possible positive diagnostic rate when s-NDDs children undergo trio-WES testing, resulting in some children being unable to obtain timely genetic diagnosis. Summary of the Invention

[0004] In view of this, it is necessary to provide a method, system, electronic device, and storage medium for constructing a neurodevelopmental disorder prediction model, which can at least overcome one of the above defects.

[0005] In a first aspect, an embodiment of the present application provides a method for constructing a neurodevelopmental disorder prediction model, which is applied to predicting the probability of a positive result in the third-generation whole exome genome sequencing examination of children with syndromic neurodevelopmental disorders. The neurodevelopmental disorder prediction method includes:

[0006] Obtain the clinical information of the child, where the clinical information includes the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain developmental deformity, impaired hearing / vision function, and related ear and eye developmental deformities;

[0007] Apply logistic regression analysis to screen out the predictors in the clinical information that are significantly correlated with the positive diagnosis rate of trio whole-exome genome sequencing;

[0008] Based on the predictors, construct a binary logistic regression-based neurodevelopmental disorder prediction model;

[0009] Use the data of the internal validation group and the external validation group to validate the neurodevelopmental disorder prediction model to evaluate the prediction efficacy of the neurodevelopmental disorder prediction model.

[0010] In one embodiment, the predictors include:

[0011] The severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, head circumference deformity, and brain developmental deformity.

[0012] In one embodiment, the complexity of the syndromic neurodevelopmental disorder includes whether the patient has at least two of autism spectrum disorder, epilepsy, or attention deficit hyperactivity disorder.

[0013] In one embodiment, the output of the neurodevelopmental disorder prediction model is the probability of obtaining a positive diagnosis after the child undergoes trio whole-exome genome sequencing.

[0014] In one embodiment, the prediction model shows the probability of the child obtaining a positive diagnosis of the trio whole-exome genome sequencing through the score distribution of the nomogram.

[0015] In one embodiment, the neurodevelopmental disorder prediction model divides the children into a high positive diagnosis probability group and a low positive diagnosis probability group based on the optimal cut-off point determined by the Youden index.

[0016] In one embodiment, the method includes a follow-up plan for the enrolled children, and the follow-up plan includes:

[0017] After the child undergoes trio whole-exome genome sequencing, obtain the test results of the child, and modify the neurodevelopmental disorder prediction model according to the test results, where the test results include positive diagnosis and negative diagnosis.

[0018] In a second aspect, an embodiment of the present application provides a neurodevelopmental disorder prediction model construction system, which is applied to implement the neurodevelopmental disorder prediction model construction method as described in the first aspect. The system includes:

[0019] An information acquisition module for acquiring the clinical information of a child, where the clinical information includes the severity of developmental delay / intellectual disability, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain developmental deformity, impaired hearing / vision function, and related ear and eye developmental deformities;

[0020] An information screening module for screening out predictors significantly related to the positive diagnosis rate of trio whole exome genomic sequencing from the clinical information by applying logistic regression analysis;

[0021] A model construction module for constructing a neurodevelopmental disorder prediction model based on the predictors by building a binary logistic regression;

[0022] A model verification module for verifying the neurodevelopmental disorder prediction model by applying the data of the internal verification group and the external verification group to evaluate the prediction efficacy of the neurodevelopmental disorder prediction model.

[0023] In a third aspect, an embodiment of the present application provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein, when the processor is configured to execute the instructions, the neurodevelopmental disorder prediction model construction method described in the first aspect is implemented.

[0024] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, including instructions for instructing a device to execute the neurodevelopmental disorder prediction model construction method described in the first aspect.

[0025] The neurodevelopmental disorder prediction model construction method, system, electronic device, and storage medium provided by the embodiments of the present application can accurately evaluate the probability of a child obtaining a positive diagnosis after receiving trio-WES testing by integrating the multi-dimensional clinical information of the child, including significantly related factors such as the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, and head circumference deformity, combining logistic regression analysis to screen predictors and construct a binary logistic regression model, significantly improving the prediction efficacy, providing clear diagnostic decision support for clinicians, and optimizing the diagnosis and treatment process and resource allocation of children. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic flowchart of a neurodevelopmental disorder prediction model construction method provided by an embodiment of the present application.

[0027] Figure 2 It is a schematic diagram of the performance of a training group provided by an embodiment of the present application.

[0028] Figure 3 It is a schematic diagram of the performance of an internal verification group provided by an embodiment of the present application.

[0029] Figure 4 It is a schematic diagram of the performance of the external verification group provided by an embodiment of the present application.

[0030] Figure 5 It is a schematic diagram of the efficacy evaluation provided by an embodiment of the present application.

[0031] Figure 6 It is a schematic diagram of the nomogram provided by an embodiment of the present application.

[0032] Figure 7 It is a schematic diagram of the system module for constructing a neurodevelopmental disorder prediction model provided by an embodiment of the present application.

[0033] Figure 8 A schematic diagram of an electronic device provided by an embodiment of the present application.

[0034] Description of main component symbols

[0035] Neurodevelopmental disorder prediction model construction system 10

[0036] Information acquisition module 11

[0037] Information screening module 12

[0038] Model construction module 13

[0039] Model verification module 14

[0040] Electronic device 20

[0041] Processor 21

[0042] Memory 22

[0043] Method steps S100 - S400 Detailed implementation manners

[0044] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments.

[0045] It should be noted that "at least one" in the embodiments of the present application means one or more, and multiple means two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application.

[0046] It should be noted that in the embodiments of the present application, terms such as "first" and "second" are only used for the purpose of distinguishing descriptions, and should not be construed as indicating or implying relative importance, nor as indicating or implying an order. Features defined with "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to mean for example, illustration or explanation. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0047] Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0048] Neurodevelopmental disorder syndromes (s-NDDs) are a group of neurological developmental diseases with complex etiologies and diverse clinical manifestations. They mainly affect the neurological development of children and usually manifest as disorders in multiple aspects such as cognition, language, and movement. These disorders may be accompanied by other mental or neurological symptoms, such as autism spectrum disorder (ASD), epilepsy (EP), and attention deficit hyperactivity disorder (ADHD). The etiology of s-NDDs has not been fully clarified, and the manifestations of the diseases show a high degree of heterogeneity, making diagnosis and treatment more difficult. Early diagnosis of these diseases is crucial because early intervention has a significant impact on improving the prognosis and quality of life of children. With the development of genomics, especially the application of third-generation genomic sequencing technology (trio-WES), the etiology of some s-NDDs patients has been clarified, but the overall diagnostic rate is still low, making it difficult to meet the clinical demand for early and accurate diagnosis.

[0049] Currently, although the trio-WES technology has made significant progress in the diagnosis of some diseases, its application in s-NDDs still faces challenges. On the one hand, due to the high phenotypic diversity of s-NDDs, the symptoms of many children are atypical, making it difficult to determine positive results in genetic testing. On the other hand, the cost of trio-WES is relatively high, and the detection process usually takes a long time, which limits its popularization in clinical applications. Therefore, the lack of a prediction model based on clinical phenotypic factors to estimate the possible positive diagnostic rate when s-NDDs children undergo trio-WES testing results in some children being unable to obtain timely genetic diagnosis.

[0050] Therefore, the embodiments of the present application provide a method, a system, an electronic device, and a storage medium for constructing a neurodevelopmental disorder prediction model. By integrating the multi-dimensional clinical information of children, including significantly relevant factors such as the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, and head circumference deformity, and combining logistic regression analysis to screen for predictive factors and construct a binary logistic regression model, it is possible to accurately evaluate the probability of a child obtaining a positive diagnosis after trio-WES testing, significantly improve the prediction efficiency, provide clear diagnostic decision support for clinicians, and optimize the diagnosis and treatment process and resource allocation for children.

[0051] The following will describe in detail some embodiments of the application with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0052] Figure 1 It is a schematic flowchart of a method for constructing a neurodevelopmental disorder prediction model provided by an embodiment of the present application. As Figure 1 shown, the method for constructing a neurodevelopmental disorder prediction model includes at least the following steps: S100: Obtain the clinical information of the child; S200: Apply logistic regression analysis to screen out the predictive factors in the clinical information that are significantly correlated with the positive diagnosis rate of trio whole exome sequencing; S300: Based on the predictive factors, construct a neurodevelopmental disorder prediction model using binary logistic regression; S400: Use the data of the internal validation group and the external validation group to verify the neurodevelopmental disorder prediction model to evaluate the prediction efficiency of the neurodevelopmental disorder prediction model.

[0053] S100: Obtain the clinical information of the child.

[0054] In the embodiments of the present application, the method for constructing a neurodevelopmental disorder prediction model includes obtaining the clinical information of the child in step S100.

[0055] Specifically, the clinical information includes the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain development deformity, hearing / vision function impairment, and related ear and eye development deformities.

[0056] In the embodiments of the present application, GDD (Global Developmental Delay) and ID (Intellectual Disability) are used as one of the predictive factors, and the severity is judged by evaluating the developmental domain scores of children under 5 years old. Specifically, if the scores of a child in at least two developmental domains are lower than 35 points, it is defined as severe to extremely severe GDD; if the scores are higher than 35 points, it is defined as mild to moderate GDD. For ID patients over 5 years old, if their IQ scores are lower than 40 points, it is defined as severe to extremely severe ID, and vice versa, it is defined as mild to moderate ID.

[0057] In the embodiments of the present application, the complexity of s-NDDs In this embodiment, the complexity of s-NDDs (neurodevelopmental disorders with combined phenotypes) is used as another predictor. According to the diagnostic criteria of DSM-V and ILAE, the enrolled children are examined for ASD (autism spectrum disorder), ADHD (attention deficit hyperactivity disorder), and EP (epilepsy). If a child has two or more combined phenotypes of neurodevelopmental disorders, it is defined as complex s-NDDs; if there is only one disorder phenotype, it is defined as simple s-NDDs.

[0058] In the embodiments of the present application, the co-occurrence of ASD, EP, and ADHD In this embodiment, the co-occurrence of ASD, EP, and ADHD in children is used as a predictor. According to the diagnostic criteria of DSM-V and ILAE, it is determined whether a child has neurodevelopmental disorders such as ASD, EP, or ADHD. If a child meets the diagnosis of any of the above disorders, it is defined as having ASD, EP, or ADHD; otherwise, it is defined as not having ASD, EP, or ADHD.

[0059] In the embodiments of the present application, the diagnostic criteria for head circumference malformation In this embodiment, whether a child has a head circumference malformation (such as microcephaly or macrocephaly) is used as an important predictor. According to the criteria of HPO (Human Phenotype Ontology), if a child has microcephaly or macrocephaly, it is defined as having a head circumference malformation; otherwise, it is defined as not having a head circumference malformation.

[0060] In the embodiments of the present application, the diagnostic criteria for brain development malformation In this embodiment, the diagnosis of brain development malformation is based on the results of cranial MR (magnetic resonance) examination. If cranial MR shows developmental malformations such as cortical dysplasia, it is defined as having a brain development malformation; otherwise, it is defined as not having a brain development malformation.

[0061] In the embodiments of the present application, impaired audiovisual function and ear-eye developmental malformations In this embodiment, whether a child has impaired audiovisual function or ear-eye developmental malformations is used as a predictor. Through ABR (auditory brainstem response) and VEP (visual evoked potential) examinations, if the results show impaired hearing or vision function (for example, the auditory response threshold is greater than 30 BnHL, or the P100 latency is prolonged), it is defined as having impaired audiovisual function. In addition, if there are abnormalities in the development of the ears and eyes (such as ear-eye developmental malformations), it is defined as having ear-eye developmental malformations.

[0062] In the embodiments of the present application, the outcome indicator of the present application for the trio-WES diagnosis result is the pathogenicity level of gene mutations detected by trio-WES (whole exome sequencing of three generations of the family). According to the 2015 ACMG guidelines, the test results are classified into pathogenic, likely pathogenic, variants of uncertain significance, and benign, etc. If the test result is classified as pathogenic or likely pathogenic, it is defined as trio-WES positive (+); if classified as variants of uncertain significance or benign, it is defined as trio-WES negative (-).

[0063] It can be understood that by collecting multi-dimensional clinical information, the basic situation of the child can be comprehensively grasped, and an accurate data basis can be provided for the construction of the subsequent prediction model.

[0064] Please refer to Table 1 together. Table 1 is a statistical table of the base characteristics and related variables.

[0065]

[0066]

[0067]

[0068] Table 1 Statistical table of base characteristics and related variables

[0069] It can be understood that Table 1 shows the statistical analysis of the base characteristics of the neurodevelopmental disorder prediction model and each group of data. The data in the table come from the training group, the internal validation group, and the external validation group respectively, covering demographic data, predictors, and the final trio-WES diagnosis results. By comparing and analyzing the cases in each group, the stability and prediction efficacy of the model can be evaluated. In terms of demographic data, the table lists the number of cases, gender distribution, visit date, and age at which trio-WES was performed in each group. The difference in the gender ratio among the three groups did not reach a significant level (P value > 0.05). In terms of the visit date, there were slight differences in the visit times of the training group, the internal validation group, and the external validation group, and there were no significant differences in the age at which each group received genetic testing (P value > 0.05). In terms of the case source, the data come from different hospitals respectively, indicating the wide applicability of the model in different medical environments.

[0070] In terms of predictors, Table 1 lists information including the severity of developmental delay / intellectual disability, the degree of complication combination (such as whether combined with autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, etc.), head circumference deformity, brain developmental deformity, vision / hearing impairment, etc. Data analysis shows that there are certain differences in the distribution of these predictors among groups, but no significant differences are found in all characteristics (for example, the severity of GDD / ID and the degree of complication combination do not vary much among groups). In particular, complications such as autism spectrum disorder (ASD) and epilepsy (EP) show a relatively consistent trend in the incidence rates among different groups, indicating that these factors may play an important role in the diagnosis of neurodevelopmental disorders.

[0071] Finally, in terms of the trio-WES diagnosis results, the table summarizes the proportions of children who received positive or negative diagnoses after undergoing trio whole-exome genome sequencing in each group. Although the proportions of children with positive diagnoses vary among different groups (for example, the positive diagnosis rate in the training group is 50.3%, and in the internal validation group is 41.2%), the P-value shows that the differences among groups are not significant. This indicates that, despite the differences in the genetic backgrounds and clinical characteristics of cases in each group, the diagnostic effectiveness of the neurodevelopmental disorder prediction model is relatively stable among different groups. That is, the data in Table 1 provide an important basis for the optimization of the neurodevelopmental disorder prediction model, especially in identifying key predictors and evaluating the generalization ability of the model.

[0072] S200: Apply logistic regression analysis to screen out predictors in the clinical information that are significantly correlated with the positive diagnosis rate of trio whole-exome genome sequencing.

[0073] In the embodiment of the present application, the method for constructing a neurodevelopmental disorder prediction model includes, in step S200, applying logistic regression analysis to screen out predictors in the clinical information that are significantly correlated with the positive diagnosis rate of trio whole-exome genome sequencing.

[0074] In this embodiment, in order to screen out the most clinically significant factors from a series of candidate predictors and construct an effective prediction model, a traditional binary Logistic regression method of first single and then multiple is adopted, and the training group data is combined for screening predictors. The specific steps are as follows:

[0075] First, in the training group data, apply univariate Logistic regression analysis to screen all possible predictors. The relationship between each predictor and the outcome indicator (such as trio-WES positive / negative) is evaluated by calculating the P-value. If the P-value of a certain predictor is less than 0.05, it indicates that there is a significant statistical association between this factor and the outcome indicator, and it is considered a meaningful predictor. All these factors with significance (P < 0.05) will enter the subsequent multivariate analysis step.

[0076] On the basis of single-factor screening, all significant predictors were included together in a multivariable Logistic regression model for further analysis. The independence of these predictors from the outcome measures was evaluated through the multivariable regression model. Applying the criterion of P < 0.05 again, factors that made independent contributions to the prediction results were selected. These factors will serve as the final predictors in the subsequent construction of the nomogram model.

[0077] To ensure that the selected predictors can function independently in the model, collinearity analysis needs to be performed on them. Collinearity analysis detects whether there is a high correlation between each predictor by calculating the variance inflation factor (VIF) and tolerance. If the VIF values of some predictors are high (usually VIF > 10), or the tolerance is low (usually tolerance < 0.1), it indicates that there may be multicollinearity between these factors, which may affect the stability and accuracy of the regression analysis. In this case, some highly correlated factors can be considered for removal or other appropriate treatments to improve the independence between the predictors.

[0078] The final predictors selected through the above steps will be incorporated into the construction process of the nomogram prediction model. The final model will provide an efficient and accurate binary classification prediction tool for clinical practice based on these reliable and non-collinear predictors, combined with the calculation results of the Logistic regression model.

[0079] Specifically, through logistic regression analysis, the most diagnostically valuable predictors can be selected based on the correlation between clinical information and the results of trio whole-exome sequencing (trio-WES). These factors play an important role in determining whether a child has a hereditary neurodevelopmental disorder.

[0080] It can be understood that this process is an in-depth exploration of clinical data, identifying important variables related to genetic pathological diagnosis through statistical methods, and providing a scientific basis for the subsequent prediction model.

[0081] S300: Construct a neurodevelopmental disorder prediction model based on the predictors using binary logistic regression.

[0082] In the embodiment of the present application, the method for constructing a neurodevelopmental disorder prediction model includes constructing a neurodevelopmental disorder prediction model based on the predictors using binary logistic regression in step S300.

[0083] In the embodiments of the present application, the research outcome index is a binary variable, that is, the diagnostic result (positive or negative) of trio-WES. Therefore, a binary classification Logistic regression model is used to construct the prediction model. The inputs of the model include multiple predictors, such as the severity of GDD / ID, the complexity of s-NDDs, the co-occurrence of ASD, EP, and ADHD, microcephaly, brain dysplasia, impaired audiovisual function, and ear and eye dysplasia, etc. Through the training of the Logistic regression model and combining various features, a model that can be used to predict the positive and negative of trio-WES is finally obtained.

[0084] In the embodiments of the present application, the patient data is first randomly divided into a training group and a validation group. The trained Logistic regression prediction model is applied to the training group cohort to evaluate its prediction efficacy. To evaluate the goodness of fit and prediction effect of the model, a calibration curve combined with the Hosmer-Lemeshow (H-L) test and the ROC curve are used for evaluation. The calibration curve is used to detect the degree of agreement between the predicted probability of the model and the actual result, the H-L test is used for statistical testing of the goodness of fit of the model, and the ROC curve is used to evaluate the sensitivity and specificity of the model to help determine the optimal cut-off value of the model.

[0085] In the embodiments of the present application, the Youden's Index is calculated using the training group cohort to determine the optimal cut-off value of the ROC curve. The score of the prediction model corresponding to this optimal cut-off value is called the nomoScore. Through the nomoScore, the internal independent validation group is divided into a high positive trio-WES diagnostic prediction rate group and a low positive trio-WES diagnostic prediction rate group. Then, methods such as the calibration curve, the H-L test, and the ROC curve are used to evaluate the goodness of fit and prediction efficacy of the prediction model in the internal independent validation group. Through these evaluations, it is verified whether the model has good generalization ability and prediction accuracy.

[0086] In the embodiments of the present application, the patient data of the external independent validation group is further applied to the prediction model. Similar to the internal independent group validation, the patients in the external validation group are first divided into a high positive trio-WES diagnostic prediction rate group and a low positive trio-WES diagnostic prediction rate group through the nomoScore. Then, methods such as the calibration curve, the H-L test, and the ROC curve are used to evaluate the goodness of fit and prediction efficacy of the prediction model in the external independent validation group. This step can further verify the stability and generalization performance of the model in different datasets, thereby ensuring its reliability and accuracy in clinical applications.

[0087] Specifically, according to the predictors selected in S200, a binary logistic regression algorithm is used to construct a neurodevelopmental disorder prediction model, which can predict whether a child is likely to obtain a positive diagnosis through trio whole-exome sequencing (trio-WES) based on different clinical characteristics.

[0088] It can be understood that this process is the core step of the model, and a quantitative model that can judge whether a child has a hereditary neurodevelopmental disorder is established through the logistic regression algorithm.

[0089] S400: Validate the neurodevelopmental disorder prediction model using the data of the internal validation group and the external validation group to evaluate the prediction efficacy of the neurodevelopmental disorder prediction model.

[0090] In the embodiment of the present application, the method for constructing a neurodevelopmental disorder prediction model includes validating the neurodevelopmental disorder prediction model using the data of the internal validation group and the external validation group in step S400 to evaluate the prediction efficacy of the neurodevelopmental disorder prediction model.

[0091] Specifically, by applying the prediction model in different data sets (internal validation group and external validation group), comparing the differences between the prediction results and the actual results of the model, and evaluating performance indicators such as the accuracy, sensitivity, and specificity of the model, the reliability and generalization ability of the model are verified.

[0092] Please refer to Figures 2 to 4 . Figure 2 It is a schematic diagram of the performance of the training group provided by an embodiment of the present application. Figure 3 It is a schematic diagram of the performance of the internal validation group provided by an embodiment of the present application. Figure 4 It is a schematic diagram of the performance of the external validation group provided by an embodiment of the present application.

[0093] As Figures 2 to 4 shown, evaluate the prediction efficacy of the model based on the training group. According to Figure 2 the results, the model has a good fit and an ideal prediction efficacy. The AUC value is 0.853, indicating that the model has strong discrimination ability. Further, the Youden index is calculated based on the training group cohort to be 0.585. Based on this standard, the nomoScore corresponding to the optimal cut-off value of the ROC is determined to be 214 points. Through this score, the internal / external independent validation groups are respectively divided into high-positive and low-positive trio-WES diagnostic probability groups.

[0094] To further verify the efficacy of the model in the validation group, a calibration curve and an ROC curve are used to evaluate it. Figure 3 and Figure 4 show the performance of the model in the internal and external independent validation groups. From Figure 3 andFigure 4 It can be seen that the model shows excellent goodness of fit and prediction efficacy in both the internal validation group and the external validation group, with AUC values of 0.902 and 0.927 respectively, further verifying that the model has good generalization ability and reliability. This indicates that the constructed neurodevelopmental disorder prediction model not only shows good prediction performance in the training group, but also maintains high accuracy and stability in the independent validation group.

[0095] Please refer to Figure 5 , Figure 5 which is a schematic diagram of efficacy evaluation provided by an embodiment of the present application. The Sankey diagram shows the main efficacy indicators of the neurodevelopmental disorder prediction model in the training group, internal validation group, and external validation group, including sensitivity, specificity, accuracy, and precision.

[0096] From Figure 5 it can be seen that in the training group, the sensitivity of the model is 65.85%, the specificity is 92.59%, the accuracy is 79.14%, and the precision is 90.00%. In the internal validation group, the sensitivity of the model is 76.19%, the specificity is 91.67%, the accuracy is 85.29%, and the precision is 86.49%. In the external validation group, the sensitivity is 66.67%, the specificity is 100.00%, the accuracy is 85.57%, and the precision is 100.00%.

[0097] These indicators show that the model demonstrates excellent prediction efficacy in different validation groups. Especially in terms of specificity, the model performs outstandingly, with specificities all higher than 90%, and reaching 100% in the external validation group. The prediction model can not only effectively identify positive cases, but also show high accuracy and precision in predicting negative cases. These results indicate that the present model has high reliability in practical applications and is applicable to predicting the positive diagnosis probability of trio-WES in children with syndromic neurodevelopmental disorders.

[0098] It can be understood that this step provides strong validation support for the clinical application of the model, ensuring that the prediction model is not only effective on training data, but also has good prediction effects in actual clinical data.

[0099] In the embodiment of the present application, the predictors include the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, microcephaly, and brain dysplasia.

[0100] Specifically, through collinearity analysis, the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, the autism spectrum disorder, the tolerance of microcephaly and brain dysplasia were 0.809, 0.936, 0.895, 0.866, 0.891 respectively, and the VIF (variance inflation factor) were 1.235, 1.068, 1.118, 1.155, 1.123 respectively. These results indicate that all predictors show low collinearity statistically, and the VIF values are all less than 10, indicating weak correlations between these factors and no problem of multicollinearity, thus ensuring the stability and accuracy of the model.

[0101] It can be understood that these clinical information can help identify the specific characteristics of neurodevelopmental disorders, thus providing more precise input variables for the prediction model. The results of collinearity analysis show that the selected predictors are independent to a certain extent and have little mutual influence. Therefore, when constructing the prediction model, the diagnostic value of each factor can be effectively retained. In addition, the low VIF values also indicate that no significant multicollinearity problem will be introduced when using these factors for regression analysis, thus avoiding adverse effects on the prediction effect of the model. Therefore, the combination of these factors provides a strong basis for establishing a stable and effective prediction model for neurodevelopmental disorders.

[0102] In the embodiments of the present application, the complexity of syndromic neurodevelopmental disorders includes whether the patient has at least two of autism spectrum disorder, epilepsy or attention deficit hyperactivity disorder.

[0103] Specifically, the presence of symptoms such as ASD, EP, ADHD, etc. indicates the diversity and complexity of neurodevelopmental disorders, and the existence of complex neurodevelopmental disorders will significantly affect the clinical manifestations of children and the detection results of gene mutations.

[0104] It can be understood that these comorbidities are key indicators of neurodevelopmental disorders, which can reflect the clinical complexity of children, thus providing more influencing factors for the construction of the prediction model.

[0105] In the embodiments of the present application, the output of the neurodevelopmental disorder prediction model is the probability of obtaining a positive diagnosis after a child undergoes trio whole exome genomic sequencing.

[0106] Specifically, the output probability of the model is the possibility of a child having a hereditary neurodevelopmental disorder, which helps clinicians make more accurate decisions in actual diagnosis.

[0107] It can be understood that the prediction model can provide a quantitative diagnosis probability based on the relationship between clinical characteristics and gene mutations, thus providing a personalized detection plan for children.

[0108] In the embodiments of the present application, the prediction model shows the probability of a child obtaining a positive diagnosis from a trio whole-exome genome sequencing through the score distribution of the nomogram.

[0109] Please refer to Figure 6 , the schematic diagram of the nomogram provided by an embodiment of the present application.

[0110] Specifically, the nomogram correlates the scores of the predictive factors with the corresponding positive diagnosis probabilities, helping doctors intuitively evaluate the diagnostic risks of children based on the scores shown in the diagram.

[0111] It can be understood that, as a visualization tool, the nomogram can clearly display the diagnostic probabilities of children under different clinical characteristics, providing intuitive and effective clinical decision support.

[0112] In the embodiments of the present application, the neurodevelopmental disorder prediction model divides children into a high positive diagnosis probability group and a low positive diagnosis probability group based on the optimal cut-off point determined by the Youden index.

[0113] Specifically, the Youden index, as an important indicator for evaluating the performance of the model, maximizes the true positive rate and the true negative rate by finding the optimal cut-off point to ensure the accuracy of the diagnosis.

[0114] It can be understood that the determination of the cut-off point helps to optimize the application of the model in different clinical scenarios, helping doctors effectively distinguish high-risk and low-risk patients.

[0115] In the embodiments of the present application, the method includes a follow-up plan for the enrolled children. The follow-up plan includes obtaining the test results of the children after they receive the trio whole-exome genome sequencing, and modifying the neurodevelopmental disorder prediction model according to the test results. The test results include positive diagnoses and negative diagnoses.

[0116] Specifically, the follow-up plan updates and optimizes the prediction model by tracking the genetic test results of the children to ensure that the model can continuously provide more accurate predictions.

[0117] It can be understood that modifying the model based on the follow-up data can not only improve the prediction accuracy of the model, but also enhance the generalization ability of the model, enabling it to adapt to the clinical situations of different patients.

[0118] Figure 7 is a schematic diagram of the modules of the neurodevelopmental disorder prediction model construction system provided by an embodiment of the present application. As Figure 7 shown, the neurodevelopmental disorder prediction model construction system 10 includes at least the following parts: an information acquisition module 11, an information screening module 12, a model construction module 13, and a model verification module 14.

[0119] In the embodiment of the present application, the information acquisition module 11 is used to acquire the clinical information of the child, and the clinical information includes the severity of developmental delay / intellectual disability, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain developmental deformity, impaired hearing / vision function, and related ear and eye developmental deformities. For details, please refer to Figures 1 to 6 and its corresponding description, which will not be elaborated herein in the present application.

[0120] In the embodiment of the present application, the information screening module 12 is used to screen out the predictors significantly related to the positive diagnosis rate of trio whole-exome genome sequencing in the clinical information by applying logistic regression analysis. For details, please refer to Figures 1 to 6 and its corresponding description, which will not be elaborated herein in the present application.

[0121] In the embodiment of the present application, the model construction module 13 is used to construct a neurodevelopmental disorder prediction model based on the predictors by building a binary logistic regression. For details, please refer to Figures 1 to 6 and its corresponding description, which will not be elaborated herein in the present application.

[0122] In the embodiment of the present application, the model verification module 14 uses the data of the internal verification group and the external verification group to verify the neurodevelopmental disorder prediction model to evaluate the prediction efficacy of the neurodevelopmental disorder prediction model. For details, please refer to Figures 1 to 6 and its corresponding description, which will not be elaborated herein in the present application.

[0123] Figure 8 This is the electronic device 2O provided by an embodiment of the present application. As Figure 8 shown, the electronic device 2O at least includes the following parts: a processor 21 and a memory 22.

[0124] In the embodiment of the present application, the memory 22 is used to store the executable instructions of the processor 21, and when the processor 21 is configured to execute the instructions, it implements the neurodevelopmental disorder prediction model construction method as Figure 1 shown.

[0125] In the embodiment of the present application, a computer-readable storage medium includes instructions that direct the device to execute the neurodevelopmental disorder prediction model construction method as in the first aspect. For example, the instructions direct the device to execute the neurodevelopmental disorder prediction model construction method shown in steps S1OO to S4OO as in Figure 1 .

[0126] The program operating in the electronic device 2O according to an embodiment of the present application may be a program that controls a central processing unit (CPU) or the like to implement the functions of the above-described embodiments related to one aspect of the present invention (a program that causes a computer to function). Then, the information processed by these devices is temporarily stored in a random access memory (RAM) during its processing, and thereafter, it is stored in various ROMs such as a read only memory (Flash ROM), a hard disk drive (HDD), etc., and is read, corrected, and written by the CPU as needed.

[0127] It should be noted that a part of the electronic device 2O of the above-described embodiment can also be implemented by a computer. In this case, the program for implementing the control function can be recorded on a computer-readable recording medium, and is implemented by reading the program recorded on the recording medium into a computer system and executing it.

[0128] It should be noted that the "computer system" mentioned here refers to the computer system built into the electronic device 2O, and a computer system including hardware such as an OS and peripheral devices is adopted. In addition, the "computer-readable recording medium" refers to a removable medium such as a floppy disk, a magneto-optical disk, a ROM, a CD-ROM, etc., and a storage device such as a hard disk built into the computer system.

[0129] Moreover, the "computer-readable recording medium" may include: a medium that dynamically stores a program for a short time, such as a communication line in the case of transmitting a program via a network such as the Internet or a communication line such as a telephone line; a medium that stores a program for a fixed time, such as a volatile memory inside a computer system of a server or a client in this case. In addition, the above program may be a program for implementing a part of the above functions, and may also be a program that can implement the above functions by being combined with a program already recorded in the computer system.

[0130] In addition, the electronic device 2O in the above-described embodiment can also be implemented as an aggregate (device group) composed of a plurality of devices. Each device constituting the device group may have a part or all of the functions or function blocks of the electronic device 2O of the above-described embodiment. As the device group, it is sufficient to have all the functions or function blocks of the electronic device 2O.

[0131] It can be understood that the method, system 10, electronic device 20, and storage medium for constructing a neurodevelopmental disorder prediction model provided by the embodiments of the present application integrate multi-dimensional clinical information of children, including significantly related factors such as the severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, and cranial malformations. By combining logistic regression analysis to screen for predictors and construct a binary logistic regression model, it is possible to accurately evaluate the probability of a child obtaining a positive diagnosis after undergoing trio-WES testing, significantly improving the prediction efficiency, providing clear diagnostic decision support for clinicians, and optimizing the diagnosis and treatment process and resource allocation for children.

[0132] Those of ordinary skill in the art of the present technology should recognize that the above embodiments are merely used to illustrate the present application and are not intended to limit the present application. As long as appropriate changes and variations made to the above embodiments fall within the scope of the spirit of the present application, they fall within the scope of protection required by the present application.

Claims

1. A method for constructing a neurodevelopmental disorder prediction model, which is used to predict the probability of a positive third-generation whole exome genome sequencing test in children with syndromic neurodevelopmental disorders, characterized in that: The neurodevelopmental disorder prediction method comprises: Obtain clinical information of the child, including the severity of developmental delay / intellectual disability, complexity of syndromic neurodevelopmental disorders, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain development malformation, hearing / vision impairment, and related ear and eye development malformations; Logistic regression analysis was used to screen out the predictive factors in the clinical information that were significantly associated with the positive diagnostic rate of third-generation whole-exome genome sequencing; A neurodevelopmental disorder prediction model was constructed based on binary logistic regression according to the predictive factors; The neurodevelopmental disorder prediction model was validated using the internal validation group and the external validation group data to evaluate the predictive efficacy of the neurodevelopmental disorder prediction model.

2. The method for predicting neurodevelopmental disorders according to claim 1, characterized in that: The predictors include: The severity of developmental delay / intellectual disability, the complexity of syndromic neurodevelopmental disorders, autism spectrum disorders, head circumference deformity and brain developmental malformations.

3. The method for predicting neurodevelopmental disorders according to claim 1, characterized in that: The complexity of the syndromic neurodevelopmental disorder includes whether the patient has at least two of autism spectrum disorder, epilepsy or attention deficit hyperactivity disorder.

4. The method for predicting neurodevelopmental disorders according to claim 1, characterized in that: The output of the neurodevelopmental disorder prediction model is the probability of a child receiving a positive diagnosis after undergoing third-generation whole exome genome sequencing.

5. The method for predicting neurodevelopmental disorders according to claim 5, characterized in that: The prediction model displays the probability of a child receiving a positive diagnosis from the third-generation whole exome genome sequencing through the score distribution of the nomogram.

6. The method for constructing a neurodevelopmental disorder prediction model according to claim 1, characterized in that: The neurodevelopmental disorder prediction model divides children into a high positive diagnosis probability group and a low positive diagnosis probability group based on the optimal cutoff point determined by the Youden Index.

7. The method for constructing a neurodevelopmental disorder prediction model according to claim 1, characterized in that: The method includes a follow-up plan for the enrolled children, and the follow-up plan includes: After the child undergoes third-generation whole exome genome sequencing, the test results of the child are obtained, and the neurodevelopmental disorder prediction model is corrected according to the test results, and the test results include positive diagnosis and negative diagnosis.

8. A system for constructing a neurodevelopmental disorder prediction model, used to implement the neurodevelopmental disorder prediction model construction method according to any one of claims 1 to 7, characterized in that: The system comprises: An information acquisition module is used to obtain clinical information of the child, including the severity of developmental delay / mental disability, autism spectrum disorder, epilepsy, attention deficit hyperactivity disorder, head circumference deformity, brain development malformation, hearing / vision impairment and related ear and eye development malformations; An information screening module, used to screen out predictive factors significantly associated with the positive diagnosis rate of third-generation whole exome genome sequencing in the clinical information by using logistic regression analysis; A model building module, used for building a neurodevelopmental disorder prediction model based on binary logistic regression according to the predictive factors; The model validation module uses the internal validation group and the external validation group data to validate the neurodevelopmental disorder prediction model to evaluate the predictive efficacy of the neurodevelopmental disorder prediction model.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the method for constructing a neurodevelopmental disorder prediction model as described in any one of claims 1 to 7 when executing the instructions.

10. A computer-readable storage medium, characterized in that: The method comprises instructions for instructing a device to execute the method for constructing a neurodevelopmental disorder prediction model as described in any one of claims 1 to 7.

Images