ARTIFICIAL INTELLIGENCE-BASED META-ANALYSIS STUDY FOR AUTOMATIC CLASSIFICATION OF PREGNANCY RISK RELATED TO FATAL BIRTH OUTCOME

Abstract

ABSTRACT “ARTIFICIAL INTELLIGENCE-BASED META-ANALYSIS STUDY FOR AUTOMATIC CLASSIFICATION OF PREGNANCY RISK RELATED TO FATAL BIRTH OUTCOME” The present invention provides Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features. The method involves collecting retrospective and prospective clinical data, pre-processing to handle missing or duplicate entries, and using feature selection techniques to identify key clinical parameters. These parameters are input into multiple machine learning models, including XG-Boost, AdaBoost, SVM, Random Forest, and LightGBM, with a meta-classifier for enhanced risk prediction. Explainable AI tools such as SHAP are employed to ensure transparency and accuracy in the model’s decisions. This system improves pregnancy risk assessment, minimizes manual diagnosis errors, and provides critical insights for clinicians, focusing on liver function and demographic data to predict pregnancy complications. Figure 1

Specification

Description:TECHNICAL FIELD
[0001] The present invention relates to the field of medical science, and more particularly, the present invention relates to the Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features.
BACKGROUND ART
[0002] The following discussion of the background of the invention is intended to facilitate an understanding of the present invention. However, it should be appreciated that the discussion is not an acknowledgment or admission that any of the material referred to was published, known, or part of the common general knowledge in any jurisdiction as of the application's priority date. The details provided herein the background if belongs to any publication is taken only as a reference for describing the problems, in general terminologies or principles or both of science and technology in the associated prior art.
[0003] Cholestasis of pregnancy is due to liver problems. It slows or stops the normal flow of bile from the gallbladder. In pregnancies complicated by intrahepatic cholestasis (ICP), elevated serum total bile acids (TBA) are associated with an increased risk of poor outcomes for the baby. Although ICP generally poses minimal threats to the mother, it significantly raises the likelihood of complications for the fetus, including premature birth, meconium passage, distress, and even death. Consequently, precise and early detection of pregnancies at high risk is crucial. The most critical diagnostic marker for ICP is elevated maternal bile acid levels. Nonetheless, distinguishing these elevated levels from those caused by other liver conditions is also essential. Besides, bile acid levels are not the sole diagnostic marker for ICP; other clinical indicators such as Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST), Alkaline Phosphatase (ALP), Total Bilirubin, Gamma-glutamyl Transferase (GGT), and Prothrombin Time (PT) also play a significant role in assessing the severity of ICP.
[0004] In today's clinical environment, physicians invest a significant amount of effort in collecting, distinguishing, and evaluating all the essential medical data before reaching a conclusion. This involves reviewing clinical records, conducting examinations, reviewing laboratory findings, and administering diagnostic tests. However, this approach is resource-intensive and requires a lot of work. Moreover, manually correlating symptoms with specific parameters due to the abundance of data and the complexity of relationships can be challenging, as the diagnostic criteria can vary between patients. Identifying the interdependencies among clinical parameters and their associated complications is also a challenging task.
[0005] Effective medical data analysis is crucial for accurate diagnosis and treatment of patients. However, traditional statistical tools like SPSS, SAS, Stata, and Minitab can be inefficient when it comes to multi-parameter correlation analysis, especially with larger datasets. Moreover, these tools require extensive time and a deep understanding of mathematical relationships between parameters, leaving room for errors in diagnosis. As a result, it is imperative to develop AI-based methods that can accurately correlate multiple clinical liver function parameters and automatically classify associated risks of fetal outcomes or pregnancy complications. With the power of AI, medical professionals can confidently make informed decisions, ensuring the best possible outcomes for their patients.
[0006] Intelligent models that can accurately identify the risk of pregnancy complications related to fetal birth outcomes by analyzing clinical data on liver function and demographic information do not currently exist. Previous statistical models proposed for this purpose have proven inefficient at performing multi-parameter correlation analysis, often taking longer when dealing with big data, such as thousands of instances with hundreds of parameters. Moreover, these models only use one or two clinical parameters, thereby missing the opportunity for multiparameter analysis. The mathematical complexity of these models requires the modeler to understand the relationship between parameters before inputting them, and they may end up overfitting when applied to big data. Clinicians currently rely on manual methods and their institutions to identify the risk outcome, but with many patients being admitted, correlating various parameters to determine the best treatment procedures becomes tedious due to the enormous amount of data generated. However, we can be confident that efforts are being made to develop new intelligent models that can accurately identify the risk of pregnancy complications to help healthcare professionals provide better care to their patients.
[0007] In light of the foregoing, there is a need for Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that overcomes problems prevalent in the prior art associated with the traditionally available method or system, of the above-mentioned inventions that can be used with the presented disclosed technique with or without modification.
[0008] All publications herein are incorporated by reference to the same extent as if each individual publication or patent application were specifically and individually indicated to be incorporated by reference. Where a definition or use of a term in an incorporated reference is inconsistent or contrary to the definition of that term provided herein, the definition of that term provided herein applies, and the definition of that term in the reference does not apply.
OBJECTS OF THE INVENTION
[0009] The principal object of the present invention is to overcome the disadvantages of the prior art by providing Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features.
[0010] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that uses 30 clinical and demographic features for the analysis.
[0011] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that uses concept of machine learning is used to develop an intelligent model for pregnancy risk identification.
[0012] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that uses advanced approach of Meta classifier to find the ultimate risk without depending on an ML model.
[0013] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that adopts feature fusion strategy by integrating several ML models.
[0014] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features, wherein the correlation of the multiple clinical data related to liver function with the demographic information is analyzed, which was missing from previous clinical analysis.
[0015] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that that performs the Influential feature analysis.
[0016] We are not following the 'black box' concept of the machine learning models. Instead, we have used Explainable AI such as SHAP feature analysis to justify the feature coordination dominant clinical features with the associated outcome.
[0017] Another object of the present invention is to provide Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features that wherein the automatic hyper-parameter tuning is performed using Sklearn Tuner.
[0018] The foregoing and other objects of the present invention will become readily apparent upon further review of the following detailed description of the embodiments as illustrated in the accompanying drawings.
SUMMARY OF THE INVENTION
[0019] The present invention relates to Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features.
[0020] The proposed solution has four steps. The first phase focuses on retrospective and prospective data collection, where the retrospective data are used for the model derivation, and the prospective data will be used for model validation. Further, it is also observed that some patient data contain missing values due to a lack of information or duplicated values because of human error. In such cases, the pre-processing of those data is necessary before analysis. In the second step, to pre-process the clinical data, we plan to use mathematical algorithms and statistical models to delete the duplicate entries and substitute or impute the missing values.
[0021] The third step involves the determination of influential clinical parameters to identify the risk of pregnancy. In clinical diagnosis, not all the parameters are equally important and contribute to the outcome. Besides, keeping the handful of irrelevant parameters may result in over fitting of the outcome, which may reduce the modeling method's effectiveness and negatively impact the performance. Hence, influential clinical parameters must be determined to classify the pregnancy risk for ICP patients (Figure 1). In our protocol, we have used several influential parameter estimation processes such as Recursive Feature Elimination (RFE), Exhaustive (EXH), Best First (BF), and Greedy Step Wise (GSW). Each feature selection method estimates the parameters in a highly correlated way to the outcome variable and improves accuracy.
[0022] After the selection of the influential clinical features, the pregnancy risk related to the fetal birth is classified by considering different machine learning (ML) models such as XG-boost, Adaboost, Support Vector Machine (SVM), and Random Forest (RF) and Light Gradient Boosting Machine (lightgbm). We have proposed a meta-classifier concept where the extracted influential clinical features are passed to the individual ML models (say Model 1, Model 2, ..., Model N) to predict the risk (Figure 1) independently. The prediction probability of the previous classifiers is passed to the meta classifier, which finally predicts the ultimate risk. The institution behind using layers of models is to make the model learn the training features appropriately and more refined.
[0023] One of the innovative features of our development is the application of Explainable AI to justify the accuracy, fairness, and transparency of the model related to feature analysis. By applying the SHAP (SHapley Additive exPlanations) tool, we can explain the features' importance and contribution to the classification outcome. After selecting the best model, a cross-validation study is performed to verify the robustness of the model. After obtaining the desired accuracy, the prospective data will be tested to verify the model's generality for patients.
[0024] While the invention has been described and shown with reference to the preferred embodiment, it will be apparent that variations might be possible that would fall within the scope of the present invention.
BRIEF DESCRIPTION OF DRAWINGS
[0025] So that the manner in which the above-recited features of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may have been referred by embodiments, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.
[0026] These and other features, benefits, and advantages of the present invention will become apparent by reference to the following text figure, with like reference numbers referring to like structures across the views, wherein:
[0027] Figure 1: Overview of the proposed model related to pregnancy risk identification related to fetal birth.
DETAILED DESCRIPTION OF THE INVENTION
[0028] While the present invention is described herein by way of example using embodiments and illustrative drawings, those skilled in the art will recognize that the invention is not limited to the embodiments of drawing or drawings described and are not intended to represent the scale of the various components. Further, some components that may form a part of the invention may not be illustrated in certain figures, for ease of illustration, and such omissions do not limit the embodiments outlined in any way. It should be understood that the drawings and the detailed description thereto are not intended to limit the invention to the particular form disclosed, but on the contrary, the invention is to cover all modifications, equivalents, and alternatives falling within the scope of the present invention as defined by the appended claim.
[0029] As used throughout this description, the word "may" is used in a permissive sense (i.e. meaning having the potential to), rather than the mandatory sense, (i.e. meaning must). Further, the words "a" or "an" mean "at least one" and the word "plurality" means "one or more" unless otherwise mentioned. Furthermore, the terminology and phraseology used herein are solely used for descriptive purposes and should not be construed as limiting in scope. Language such as "including," "comprising," "having," "containing," or "involving," and variations thereof, is intended to be broad and encompass the subject matter listed thereafter, equivalents, and additional subject matter not recited, and is not intended to exclude other additives, components, integers, or steps. Likewise, the term "comprising" is considered synonymous with the terms "including" or "containing" for applicable legal purposes. Any discussion of documents, acts, materials, devices, articles, and the like are included in the specification solely for the purpose of providing a context for the present invention. It is not suggested or represented that any or all these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention.
[0030] In this disclosure, whenever a composition or an element or a group of elements is preceded with the transitional phrase "comprising", it is understood that we also contemplate the same composition, element, or group of elements with transitional phrases "consisting of", "consisting", "selected from the group of consisting of, "including", or "is" preceding the recitation of the composition, element or group of elements and vice versa.
[0031] The present invention is described hereinafter by various embodiments with reference to the accompanying drawing, wherein reference numerals used in the accompanying drawing correspond to the like elements throughout the description. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiment set forth herein. Rather, the embodiment is provided so that this disclosure will be thorough and complete and will fully convey the scope of the invention to those skilled in the art. In the following detailed description, numeric values and ranges are provided for various aspects of the implementations described. These values and ranges are to be treated as examples only and are not intended to limit the scope of the claims. In addition, several materials are identified as suitable for various facets of the implementations. These materials are to be treated as exemplary and are not intended to limit the scope of the invention.
[0032] The present invention relates to Artificial intelligence-based meta-analysis study for automatic classification of pregnancy risk related to fatal birth outcome by correlating liver functional clinical and demographic features.
[0033] The proposed solution has four steps. The first phase focuses on retrospective and prospective data collection, where the retrospective data are used for the model derivation, and the prospective data will be used for model validation. Further, it is also observed that some patient data contain missing values due to a lack of information or duplicated values because of human error. In such cases, the pre-processing of those data is necessary before analysis. In the second step, to pre-process the clinical data, we plan to use mathematical algorithms and statistical models to delete the duplicate entries and substitute or impute the missing values.
[0034] The third step involves the determination of influential clinical parameters to identify the risk of pregnancy. In clinical diagnosis, not all the parameters are equally important and contribute to the outcome. Besides, keeping the handful of irrelevant parameters may result in over fitting of the outcome, which may reduce the modeling method's effectiveness and negatively impact the performance. Hence, influential clinical parameters must be determined to classify the pregnancy risk for ICP patients (Figure 1). In our protocol, we have used several influential parameter estimation processes such as Recursive Feature Elimination (RFE), Exhaustive (EXH), Best First (BF), and Greedy Step Wise (GSW). Each feature selection method estimates the parameters in a highly correlated way to the outcome variable and improves accuracy.
[0035] After the selection of the influential clinical features, the pregnancy risk related to the fetal birth is classified by considering different machine learning (ML) models such as XG-boost, Adaboost, Support Vector Machine (SVM), and Random Forest (RF) and Light Gradient Boosting Machine (lightgbm). We have proposed a meta-classifier concept where the extracted influential clinical features are passed to the individual ML models (say Model 1, Model 2, ..., Model N) to predict the risk (Figure 1) independently. The prediction probability of the previous classifiers is passed to the meta classifier, which finally predicts the ultimate risk. The institution behind using layers of models is to make the model learn the training features appropriately and more refined.
[0036] One of the innovative features of our development is the application of Explainable AI to justify the accuracy, fairness, and transparency of the model related to feature analysis. By applying the SHAP (SHapley Additive exPlanations) tool, we can explain the features' importance and contribution to the classification outcome. After selecting the best model, a cross-validation study is performed to verify the robustness of the model. After obtaining the desired accuracy, the prospective data will be tested to verify the model's generality for patients.
[0037] The developed model is entirely automatic and intelligent to predict the pregnancy risk related to fatal birth outcomes by correlating multiple liver function clinical parameters and demographic data. The design of an automatic risk classification model using AI can significantly impact fetal outcomes or pregnancy complications by minimizing physicians' burden of manual diagnosis, which is tedious and error-prone. Apart from this, the identification of influential features will help the clinicians for further follow-up of only those significant features. Compared to the traditional manual or statistical method, the robust intelligent pregnancy risk classification model produces the output within the stipulated time and precise accuracy.
[0038] This study aims to develop integrated and compatible software that can easily be installed in hospital computers and provide efficient analysis after inputting the respective clinical data. The developed system does not need any experts for further study. Any physician can input the respective clinical parameters, and the designed model will dominate clinical features and the associated risk.
[0039] An alternative solution is to develop a pregnancy risk classification model using individual machine-learning strategies instead of a Meta classifier. The entire data set can be trained and tested using individual machine learning models such as XG-boost, Adaboost, SVM, RF, and light GBM with individual hyper parameters tuning. Although we can benefit from the particular model's behavior towards the clinical features, the model development phase might be time-consuming, resource-exhaustive, and have a higher chance of bias towards parameter settings.
- The concept of machine learning is used to develop an intelligent model for identifying pregnancy risk for ICP suspect patients. Currently, there is no work related to AI regarding the classification of pregnancy risk and the outcome of childbirth in the case of intrahepatic cholestasis (ICP) patients. The advanced approach of Meta classifier is used to find the ultimate risk without depending on an ML model
- The feature fusion strategy is adopted by integrating several ML models
- Correlation of the multiple clinical data such as bile acid, Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST), Alkaline Phosphatase (ALP), Total Bilirubin, Gamma-glutamyl Transferase (GGT), and Prothrombin Time (PT) related to liver function with the demographic information (age, sex, etc) is analyzed, which was missing previous clinical analysis
- Influential feature analysis is performed.
- We are not following the 'black box' concept of the machine learning models. Instead, we have used Explainable AI such as SHAP feature analysis to justify the feature coordination dominant clinical features with the associated outcome
- Automatic hyper-parameter tuning is performed using Sklearn Tuner.
[0040] The application of the approach is the classification of pregnancy risk and the outcome of fatal birth in the case of ICP patients. In addition, the development applies to finding out the liver function, the existence of diabetes, and the severity of ICP in the case of the mother. In addition, the chances of premature birth and fetal health can also be diagnosed based on the risk outcome of the model.
[0041] Various modifications to these embodiments are apparent to those skilled in the art from the description and the accompanying drawings. The principles associated with the various embodiments described herein may be applied to other embodiments. Therefore, the description is not intended to be limited to the 5 embodiments shown along with the accompanying drawings but is to be providing the broadest scope consistent with the principles and the novel and inventive features disclosed or suggested herein. Accordingly, the invention is anticipated to hold on to all other such alternatives, modifications, and variations that fall within the scope of the present invention and appended claims. , Claims:CLAIMS
We Claim:
1) A method for classifying pregnancy risk related to fetal birth outcomes, the method comprising:
Collecting retrospective and prospective data to derive and validate a model, the retrospective data being used for model training and the prospective data for model validation;
Pre-processing clinical data using mathematical algorithms and statistical models to remove duplicate entries and impute missing values;
Identifying influential clinical parameters using feature selection methods including Recursive Feature Elimination (RFE), Exhaustive Search (EXH), Best First (BF), and Greedy Stepwise (GSW) to determine the most relevant parameters for classifying pregnancy risk;
Classifying pregnancy risk using machine learning models such as XG-Boost, AdaBoost, Support Vector Machine (SVM), Random Forest (RF), and Light Gradient Boosting Machine (LightGBM); and
Utilizing a meta-classifier, wherein the influential clinical features are fed into individual machine learning models to predict the risk, and the prediction probabilities from these models are passed to the meta-classifier for final risk prediction.
2) The method as claimed in claim 1, wherein the method further comprising the application of Explainable AI (XAI) techniques, particularly SHapley Additive exPlanations (SHAP), to explain the contribution and importance of the clinical features to the classification outcome.
3) The method as claimed in claim 1, wherein the clinical parameters analyzed include liver function markers such as bile acid, Alanine Aminotransferase (ALT), Aspartate Aminotransferase (AST), Alkaline Phosphatase (ALP), Total Bilirubin, Gamma-glutamyl Transferase (GGT), and Prothrombin Time (PT), as well as demographic information such as age and sex.
4) The method as claimed in claim 1, wherein the automatic hyper-parameter tuning of the machine learning models is performed using Sklearn Tuner.
5) A system for classifying pregnancy risk related to fatal birth outcomes, the system comprising:
A data collection module configured to collect retrospective and prospective clinical data for model training and validation;
A pre-processing module configured to cleanse and impute missing data using statistical and algorithmic techniques;
A feature selection module to identify influential clinical parameters using methods such as RFE, EXH, BF, and GSW;
A machine learning classification module comprising multiple models including XG-Boost, AdaBoost, SVM, RF, and LightGBM, integrated with a meta-classifier to predict pregnancy risk;
An Explainable AI module to analyze feature importance using SHAP and improve model transparency and fairness.

Documents