Role of artificial intelligence in the prediction of small-for-gestational-age birth weight  by gestational week

Role of artificial intelligence in the prediction of small-for-gestational-age birth weight by gestational week | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Role of artificial intelligence in the prediction of small-for-gestational-age birth weight by gestational week Zafer Bütün, Ece Akça Salık, Yeliz Kaya, Özer Çelik, Tuğba Tahta, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4850407/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Small for gestational age (SGA) is a significant concern in obstetrics, with implications for stillbirth, neonatal mortality, and long-term health outcomes. Early detection of SGA is crucial for prevention and treatment, but current methods have limitations. This study aimed to develop an artificial intelligence (AI)-based algorithm to predict SGA using sociodemographic and obstetric features during pregnancy. A total of 102 pregnant women meeting specific criteria were included in the study. The feature impact factors considered important factors for predicting SGA at birth were maternal weight, length, age, gravida, and parity. The LGBM model demonstrated the highest accuracy rate (71.4%) and AUC-ROC (62.7%) in predicting SGA, showcasing its potential for improving the prediction and treatment of SGA pregnancies. The study highlights the importance of using AI-driven methods in obstetrics to improve decision-making and patient care in high-risk pregnancy scenarios. Although AI/ML techniques have shown promise in enhancing the screening for SGA, further refinement and validation of algorithms are necessary before clinical implementation. Consistency in diagnostic criteria and quality assessment is essential for the widespread adoption of these methods in clinical settings. Small for gestational age (SGA) artificial intelligence machine learning prediction LGBM model obstetrics high-risk pregnancy Figures Figure 1 Figure 2 Introduction Small for gestational age (SGA) is when a fetus fails to reach its full growth potential due to a pathological factor, often placental dysfunction. SGA is a major cause of stillbirth, neonatal mortality, and short and long-term health issues worldwide ( 1 , 2 ). Fetuses with fetal growth restriction are at risk for perinatal morbidity and mortality, as well as poor long-term health outcomes, such as impaired neurological and cognitive development, and cardiovascular and endocrine diseases in adulthood ( 3 , 5 ). The onset of SGA is subtle and can only be diagnosed at the time of delivery, which poses great challenges for prevention and treatment, and it is also accompanied by a variety of complications that are hard to identify. For these reasons, many committees have summarized multiple methods for predicting SGA ( 4 , 5 , 6 , 7 , 8 ) and scholars worldwide generally focus on exploring changes in SGA-related detection parameters and the possible prediction of SGA ( 9 ). The use of machine learning in the medical field has evolved significantly over the last decade ( 10 , 11 , 12 ). However, there is limited evidence for its use in diagnosing intrauterine growth restriction. A meta-analysis of 10 studies on the predictive role of machine learning in identifying fetuses at risk for intrauterine growth restriction during pregnancy found that machine learning techniques were effective ( 13 ). We aimed to determine the most effective artificial intelligence-based algorithm for predicting SGA during pregnancy using sociodemographic and obstetric features. Materials-Methods The pregnant women that were admitted to the Eskisehir City Hospital Department of Obstetrics between 2022 and 2023 were retrospectively evaluated after receiving approval from the ethical committee (Ankara Medipol University Ethical Committee 08.01.2024/8). The inclusion criteria were as follows: • aged ˃18 or ˂ 40 years old • spontaneous, non-anomalous, and singleton pregnancy • delivery of a non-malformed liveborn or stillborn neonate at ≥24 weeks gestation • availability of all data regarding socio-demographic findings related to the study (smoking, hypertension, diabetes mellitus, and systemic lupus erythematosus) during the pre-conceptional period, obstetric history (gravida, parity, previous birth weight), delivery time, and baby's birth weight • absence of severe fetal chromosomal or structural abnormalities • no history of hypertension, diabetes mellitus, or systemic lupus erythematosus in previous pregnancies • no history of neonatologic disorders in previous pregnancies • no pregnancy with aneuploidy or major fetal abnormality resulting in termination, miscarriage, or fetal death before; 24 weeks gestation After selecting the eligible pregnant women for the study, they were categorized as small for gestational age (SGA) and non-SGA. SGA was defined as the birth of a neonate with a birth weight ˂10th percentile based on gender and birth week. The Shapiro-Wilk test was used to analyze the normality of the variables. Continuous variables were presented as mean ± standard deviation, and categorical variables as frequency (percentage). It was observed that the continuous variables used to compare the differences between the groups did not follow a normal distribution. Therefore, the Mann-Whitney U test was utilized to compare continuous variables that did not have a normal distribution between the groups. The Chi-Square test was used to examine the relationship between categorical variables among groups. A significance level of p˃0.05 was considered statistically significant in all analyses. The study was conducted following the principles of machine learning (ML) and utilized the Extra Trees Classifier, Average (AVG) Blender, Light Gradient Boosting Machine (LGBM) Classifier, eXtreme Gradient Boosting (XGB) Classifier, Logistic Regression, and Random Forest Classifier ML algorithms. In order to train the machine learning models, 80% of the data (pertaining to 81 pregnant individuals) was used for training, while the remaining 20% (pertaining to 21 patients) was used for testing. The accuracy, sensitivity, and specificity values of the models were assessed using confusion matrix metrics, as well as the area under the receiver operating characteristic (ROC) curve analysis graph. Cross-validation, a statistical method utilized to evaluate the performance of machine learning models, was employed. This analysis is commonly used in applied machine learning to compare and select the most suitable model for a particular predictive modeling task. The effectiveness of classification systems is often evaluated by examining the information provided in a confusion matrix, which displays actual and predicted classifications. Significant independent variables that impact the gestational age of the newborn (dependent variables) were identified using the permutation feature importance method. This method assesses the decrease in the model score when the value of a single variable is randomly shuffled. The Permutation Feature Importance Plot is presented in Figure 1. Results The demographic and obstetric findings of the 102 pregnant women in the study are shown in Table 1 , 52 women were assigned to the SGA and 50 to the non-SGA group (Table 1 ). There were no statistically significant differences between groups in terms of gravida, parity, number of nulliparous women, and type of previous birth. However, there were statistically significant differences in maternal age, smoking habits during the pregestational period, previous birth weight, and preterm delivery between the two groups (p = 0.049, 0.001, 0.001, and 0.001, respectively). Both the maternal age and previous birth weight were lower in pregnant women with SGA, and smoking habits and preterm delivery were also more common in this group. None of the pregnant women included in the study had hypertension, diabetes mellitus, or SLE during the pregestational period, but gestational hypertension and diabetes mellitus were detected in respectively 1 and 6 pregnant women with SGA during follow-up visits. Out of 102 pregnancies, 101 were liveborn babies, and 1 baby born to a pregnant woman with SGA was stillborn. Table 1 The demographic and obstetric findings of the patients included in the study. Parameters SGA (+) (52, 50.9%) SGA (-) (50, 49.1%) p Maternal age (years) 26.04 \(\:\mp\:\) 4.44 27.94 \(\:\mp\:\) 5.09 0.049 a Pre-pregnancy BMI (kg/m2) 29.15 \(\:\mp\:\) 2.69 30.05 \(\:\mp\:\) 4.45 0.357 a Smoking habits during the pre-pregnancy period (n, %) + 3 (100%) 0 (0%) 0.001 b - 49 (49.49%) 50 (50.51%) Gravida 1.75 \(\:\mp\:\) 0.97 1.84 \(\:\mp\:\) 0.96 0.547 a Parity 0.56 \(\:\mp\:\) 0.67 0.68 \(\:\mp\:\) 0.87 0.650 a Previous type of birth Nulliparous 27 (50.94%) 26 (49.06%) 0.878 b SVD 16 (48.48%) 17 (51.52%) C/S 9 (56.25%) 7 (43.75%) Previous birth weight 2895.40 \(\:\mp\:\) 445.13 3329,.8 \(\:\mp\:\) 323.60 0.001 a Preterm delivery 11 (78.5%) 3 (21.5%) 0.001 b a: Mann-Whitney U test (Data is not normal) b: Chi-Square Test As shown in Fig. 1 , maternal weight, length, age, gravida, and parity were identified as feature parameters. After training the data, the test data was used to assess the ML algorithms and it was determined that the LGBM model was the most powerful algorithm with the highest accuracy rate (71.4%) and AUC-ROC (62.7%) for predicting SGA at birth using maternal sociodemographic findings and obstetric history (Table 2 , Fig. 2 ). Table 2 Prognosis Prediction Results of Different Machine Learning Algorithms. Model Training Results Test Cross-Validation Model Name ROC-AUC Accuracy Interval ROC-AUC Accuracy Interval Random Forest Classifier 0.591 0.524 0.31–0.737 0.681 0.605 0.429–0.781 AVG Blender 0.632 0.642 0.36–0.783 0.584 0.633 0.464–0.82 Logistic Regression 0.617 0.593 0.274–0.682 0.479 0.511 0.412–0.773 LGBM Classifier 0.627 0.714 0.458–0.847 0.589 0.568 0.382–0.754 Extra Trees Classifier 0.673 0.617 0.192–0.591 0.666 0.667 0.438–0.797 XGB Classifier 0.664 0.667 0.465–0.868 0.637 0.667 0.49–0.843 AVG Blender: Average Blender, LGBM: Light Gradient Boosting Machine, XGB: eXtreme Gradient Boosting The confusion matrix revealed that 9 out of 11 cases of SGA at birth and 6 out of 10 cases of non-SGA at birth were accurately predicted by the LGBM algorithm in the training set (Table 2 ). Discussion The ability to predict intrauterine growth restriction has the potential to change the course of the disease and will help improve monitoring, allow for early identification of the condition, and prevent or reduce fetal growth retardation. Therefore, it is extremely important to identify maternal risk factors that can predict intrauterine growth restriction in the early stages of pregnancy. The purpose of this study was to create a machine-learning model to identify small for gestational age (SGA) by forecasting the probability of pregnant women developing SGA. Recent studies have suggested that clinical models that include maternal background and Doppler ultrasound in the first or second trimester have shown only moderate success ( 4 ). Based on these findings, there is potential for improvement by using artificial intelligence (AI) and machine learning techniques in prenatal risk assessment ( 12 , 14 ). Several studies have been conducted using different artificial intelligence models to detect fetal growth retardation. When additional imaging methods and parameters, such as CTG (cardiotocography), are included in the clinical data, the detection rates for SGA can increase up to 93%. Artificial intelligence and machine learning have demonstrated the potential to enhance predictions, particularly with CTG readings ( 15 ). However, there are limitations to the meta-analysis, as not all studies utilized the same input variables and samples. To establish a consensus, clinical prospective studies, such as randomized controlled trials, are necessary. Additionally, conducting more studies that compare AI/ML with traditional diagnostics could aid in enhancing the feature selection process and improving algorithm performance in clinical settings ( 13 ). Currently, artificial intelligence (AI) techniques are being used to improve the accuracy of risk assessment and predict adverse perinatal outcomes. These methods are utilized in various areas, including general pregnancy risk assessment, prenatal diagnosis, pregnancy diagnosis, hypertensive disorders of pregnancy, fetal growth, stillbirth, gestational diabetes, preterm delivery, and route of delivery. A systematic review revealed that artificial neural network (ANN) methods are the most effective AI applications for aiding the prediction of medical conditions ( 16 ). The application of AI techniques has the potential to assist obstetricians in making informed decisions. These techniques can provide valuable insights and predictive capabilities, which can be used in a number of ways. For example, they can be employed in prenatal screening and diagnosis, route of delivery prediction, postpartum care, and personalized care management. They can also be used to predict preterm delivery. By identifying potential risk factors and using trained and validated software, it is possible to detect complications at an early stage, which could then be prevented. This has been demonstrated in the case of patients with preterm delivery during pregnancy ( 17 ). The dynamic prediction model discussed in this study can identify the risk factors linked to SGA based on the gestational age of pregnant women. It uses a range of models at various gestational ages for prediction, making it straightforward, quick, and easy to implement. Additionally, it holds significant clinical importance. The study is limited by its small sample size, its retrospective nature at a single center, and the fact that it only involved Turkish patients. The study's strength is compromised by the lack of additional imaging methods, the absence of laboratory parameters, and the exclusion of additional calculation methods. Using machine learning algorithms, like the LGBM model, shows potential for improving the prediction and treatment of SGA pregnancies. This method is cheaper, simpler, and easier to access than other methods because it does not need extra imaging or laboratory methods and can offer useful information for healthcare providers to intervene sooner and enhance outcomes for mothers and babies. The results of the study emphasize the significance of utilizing AI-driven methods in obstetrics to improve decision-making and patient care in situations of high-risk pregnancy. More research and confirmation of these predictive models could help in creating individualized approaches for identifying and treating SGA pregnancies. Conclusion Our research shows that AI/ML techniques have the potential to improve the screening for SGA in a more precise and cost-effective way, ultimately enhancing pregnancy outcomes. However, further improvements and refinements to the algorithms are needed before these methods can be used in clinical settings. Additionally, it is crucial to ensure the consistency of diagnostic criteria and quality assessment before widespread adoption. Declarations Acknowledgements None Author contributions Conception and design of study: Z.B., E.A.S., Y.K; analysis of data and writing of the manuscript: Z.B., Ö.Ç., T.T., A.A.Y.; critical revision of the article for intellectual content: Z.B., E.A.S., Y.K., T.T. All authors read and approved the final manuscript. Funding None. Data availability All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author. Ethics approval and consent to participate The study was conducted in strict accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Hospital (Ankara Medipol University Ethical Committee 08.01.2024/8). All participants gave written informed consent before enrollment. Consent for publication This manuscript does not report personal data, such as individual details, so consent for publication is not applicable. Competing interests The authors declare no competing interests. References Gordijn, S.J., I.M. Beune, and W. Ganzevoort, Building consensus and standards in fetal growth restriction studies. Best Pract Res Clin Obstet Gynaecol, 2018. 49 : p. 117-126. Melamed, N., et al., FIGO (International Federation of Gynecology and Obstetrics) initiative on fetal growth: Best practice advice for screening, diagnosis, and management of fetal growth restriction. International Journal of Gynaecology and Obstetrics, 2021/03. 152 (Suppl 1). Jaddoe, V.W., et al., First trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study. BMJ, 2014. 348 : p. g14. Crovetto, F., et al., First-trimester screening for early and late small-for-gestational-age neonates using maternal serum biochemistry, blood pressure and uterine artery Doppler. Ultrasound Obstet Gynecol, 2014. 43 (1): p. 34-40. Gordijn, S.J., et al., Consensus definition of fetal growth restriction: a Delphi procedure. Ultrasound Obstet Gynecol, 2016. 48 (3): p. 333-9. Anon, ACOG Practice Bulletin No. 204: Fetal Growth Restriction. Obstetrics & Gynecology, 2019. 133 (2): p. e97-e109. Lees, C.C., et al., ISUOG Practice Guidelines: diagnosis and management of small-for-gestational-age fetus and fetal growth restriction. Ultrasound Obstet Gynecol, 2020. 56 (2): p. 298-312. Salomon, L.J., et al., ISUOG Practice Guidelines: ultrasound assessment of fetal biometry and growth. Ultrasound Obstet Gynecol, 2019. 53 (6): p. 715-723. Zheng, C., et al., Construction of prediction model for fetal growth restriction during first trimester in an Asian population. Ultrasound in Obstetrics & Gynecology, 2024. 63 (3): p. 321-330. Akazawa, M., et al., Machine learning approach for the prediction of postpartum hemorrhage in vaginal birth. Sci Rep, 2021. 11 (1): p. 22620. Cordina, M., et al., Maternal hemoglobin at 27-29 weeks' gestation and severity of pre-eclampsia. J Matern Fetal Neonatal Med, 2015. 28 (13): p. 1575-80. Javaid, M., et al., Significance of machine learning in healthcare: Features, pillars and applications. International Journal of Intelligent Networks, 2022. 3 : p. 58-73. Rescinito, R., et al., Prediction Models for Intrauterine Growth Restriction Using Artificial Intelligence and Machine Learning: A Systematic Review and Meta-Analysis. Healthcare (Basel), 2023. 11 (11). Lian, C., et al., Dynamic prediction model of fetal growth restriction based on support vector machine and logistic regression algorithm. Front Surg, 2022. 9 : p. 951908. Warmerdam, G.J.J., et al., Detection rate of fetal distress using contraction-dependent fetal heart rate variability analysis. Physiol Meas, 2018. 39 (2): p. 025008. Feduniw, S., et al., Application of Artificial Intelligence in Screening for Adverse Perinatal Outcomes-A Systematic Review. Healthcare (Basel), 2022. 10 (11). Rittenhouse, K.J., et al., Improving preterm newborn identification in low-resource settings with machine learning. PLoS One, 2019. 14 (2): p. e0198919. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4850407","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":345996800,"identity":"f65da9ca-a8e1-47b1-843f-931790454d7b","order_by":0,"name":"Zafer Bütün","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIiWNgGAWjYFACxgcMjA0gBvMxMJ+NnaAWZgOQFgmg2jQGhgQgxUy8Fh4zsBYGQlrk2w8zPq7ccbjO4HbPtwcff2yT52NmYPzwMQePT3qSmQ3PnjksYXDn7HbDGQm3DduYGZglZ27D4yyG/GOSjW1ALTdyt0nzJNxmBGphY+bFo4WN/zH7T4iWnGcgLfYEtfBIJLMxQrWwgbQkEtQiIfGYWbLxTLrkzBtpZpIz0m4ntzEzNuP1i3x/MuPHxh3W/Hw3kp9JfLC5bTu/vfngh494tGADkMQwCkbBKBgFo4ACAABPgUyNEJEWzwAAAABJRU5ErkJggg==","orcid":"","institution":"Private Clinic","correspondingAuthor":true,"prefix":"","firstName":"Zafer","middleName":"","lastName":"Bütün","suffix":""},{"id":345996801,"identity":"a54b11d1-cec1-4a21-8d30-86d9ec7caf66","order_by":1,"name":"Ece Akça Salık","email":"","orcid":"","institution":"Eskişehir City Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ece","middleName":"Akça","lastName":"Salık","suffix":""},{"id":345996802,"identity":"32eba7e5-efe9-4a18-8bb2-dc8ba04823b1","order_by":2,"name":"Yeliz Kaya","email":"","orcid":"","institution":"Eskişehir Osmangazi University","correspondingAuthor":false,"prefix":"","firstName":"Yeliz","middleName":"","lastName":"Kaya","suffix":""},{"id":345996803,"identity":"66cca208-1b88-4739-8a16-176d0b3345f6","order_by":3,"name":"Özer Çelik","email":"","orcid":"","institution":"Eskişehir Osmangazi University","correspondingAuthor":false,"prefix":"","firstName":"Özer","middleName":"","lastName":"Çelik","suffix":""},{"id":345996804,"identity":"8d89f68c-0726-4b39-862b-5a458e4a3f03","order_by":4,"name":"Tuğba Tahta","email":"","orcid":"","institution":"Ankara Medipol University","correspondingAuthor":false,"prefix":"","firstName":"Tuğba","middleName":"","lastName":"Tahta","suffix":""},{"id":345996805,"identity":"d099e36e-e7af-49be-8bb0-525ec8a6dd16","order_by":5,"name":"Arzu Altun Yavuz","email":"","orcid":"","institution":"Eskişehir Osmangazi University","correspondingAuthor":false,"prefix":"","firstName":"Arzu","middleName":"Altun","lastName":"Yavuz","suffix":""}],"badges":[],"createdAt":"2024-08-02 20:18:55","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4850407/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4850407/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":64166867,"identity":"5281121f-96fe-4f81-b10f-b24c6d2d792e","added_by":"auto","created_at":"2024-09-09 09:41:36","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":12876,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFeature Importance Chart of the variables\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e1: Weight 2: Length 3: Maternal age 4: Gravida 5: Parity\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4850407/v1/04ef4d2562f9a6a3b64c1ed8.png"},{"id":64166868,"identity":"eaaa6e8d-2d1a-47d4-a29d-8415cda23948","added_by":"auto","created_at":"2024-09-09 09:41:36","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":58669,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eReceiver operating characteristic curve graphs\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4850407/v1/f1a2c4ca1e46029fe0cae655.png"},{"id":72657772,"identity":"6ff6cd6d-110d-4500-8530-41368573b931","added_by":"auto","created_at":"2024-12-31 00:01:20","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":478643,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4850407/v1/62af383f-b0bf-43d8-a579-3a54af1fba41.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Role of artificial intelligence in the prediction of small-for-gestational-age birth weight by gestational week","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSmall for gestational age (SGA) is when a fetus fails to reach its full growth potential due to a pathological factor, often placental dysfunction. SGA is a major cause of stillbirth, neonatal mortality, and short and long-term health issues worldwide (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Fetuses with fetal growth restriction are at risk for perinatal morbidity and mortality, as well as poor long-term health outcomes, such as impaired neurological and cognitive development, and cardiovascular and endocrine diseases in adulthood (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). The onset of SGA is subtle and can only be diagnosed at the time of delivery, which poses great challenges for prevention and treatment, and it is also accompanied by a variety of complications that are hard to identify. For these reasons, many committees have summarized multiple methods for predicting SGA (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e) and scholars worldwide generally focus on exploring changes in SGA-related detection parameters and the possible prediction of SGA (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe use of machine learning in the medical field has evolved significantly over the last decade (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). However, there is limited evidence for its use in diagnosing intrauterine growth restriction. A meta-analysis of 10 studies on the predictive role of machine learning in identifying fetuses at risk for intrauterine growth restriction during pregnancy found that machine learning techniques were effective (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eWe aimed to determine the most effective artificial intelligence-based algorithm for predicting SGA during pregnancy using sociodemographic and obstetric features.\u003c/p\u003e"},{"header":"Materials-Methods","content":"\u003cp\u003eThe pregnant women that were admitted to the Eskisehir City Hospital Department of Obstetrics between 2022 and 2023 were retrospectively evaluated after receiving approval from the ethical committee (Ankara Medipol University Ethical Committee 08.01.2024/8). The inclusion criteria were as follows:\u003c/p\u003e\n\u003cp\u003e• aged ˃18 or ˂ 40 years old\u003c/p\u003e\n\u003cp\u003e• spontaneous, non-anomalous, and singleton pregnancy\u003c/p\u003e\n\u003cp\u003e• delivery of a non-malformed liveborn or stillborn neonate at ≥24 weeks gestation\u003c/p\u003e\n\u003cp\u003e• availability of all data regarding socio-demographic findings related to the study (smoking, hypertension, diabetes mellitus, and systemic lupus erythematosus) during the pre-conceptional period, obstetric history (gravida, parity, previous birth weight), delivery time, and baby's birth weight\u003c/p\u003e\n\u003cp\u003e• absence of severe fetal chromosomal or structural abnormalities\u003c/p\u003e\n\u003cp\u003e• no history of hypertension, diabetes mellitus, or systemic lupus erythematosus in previous pregnancies\u003c/p\u003e\n\u003cp\u003e• no history of neonatologic disorders in previous pregnancies\u003c/p\u003e\n\u003cp\u003e• no pregnancy with aneuploidy or major fetal abnormality resulting in termination, miscarriage, or fetal death before; 24 weeks gestation\u003c/p\u003e\n\u003cp\u003eAfter selecting the eligible pregnant women for the study, they were categorized as small for gestational age (SGA) and non-SGA. SGA was defined as the birth of a neonate with a birth weight\u0026nbsp;˂10th percentile based on gender and birth week.\u003c/p\u003e\n\u003cp\u003eThe Shapiro-Wilk test was used to analyze the normality of the variables. Continuous variables were presented as mean ± standard deviation, and categorical variables as frequency (percentage). It was observed that the continuous variables used to compare the differences between the groups did not follow a normal distribution. Therefore, the Mann-Whitney U test was utilized to compare continuous variables that did not have a normal distribution between the groups. The Chi-Square test was used to examine the relationship between categorical variables among groups. A significance level of p˃0.05 was considered statistically significant in all analyses.\u003c/p\u003e\n\u003cp\u003eThe study was conducted following the principles of machine learning (ML) and utilized the Extra Trees Classifier, Average (AVG) Blender, Light Gradient Boosting Machine (LGBM) Classifier, eXtreme Gradient Boosting (XGB) Classifier, Logistic Regression, and Random Forest Classifier ML algorithms. In order to train the machine learning models, 80% of the data (pertaining to 81 pregnant individuals) was used for training, while the remaining 20% (pertaining to 21 patients) was used for testing. The accuracy, sensitivity, and specificity values of the models were assessed using confusion matrix metrics, as well as the area under the receiver operating characteristic (ROC) curve analysis graph.\u003c/p\u003e\n\u003cp\u003eCross-validation, a statistical method utilized to evaluate the performance of machine learning models, was employed. This analysis is commonly used in applied machine learning to compare and select the most suitable model for a particular predictive modeling task. The effectiveness of classification systems is often evaluated by examining the information provided in a confusion matrix, which displays actual and predicted classifications.\u003c/p\u003e\n\u003cp\u003eSignificant independent variables that impact the gestational age of the newborn (dependent variables) were identified using the permutation feature importance method. This method assesses the decrease in the model score when the value of a single variable is randomly shuffled. The Permutation Feature Importance Plot is presented in Figure 1.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe demographic and obstetric findings of the 102 pregnant women in the study are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, 52 women were assigned to the SGA and 50 to the non-SGA group (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). There were no statistically significant differences between groups in terms of gravida, parity, number of nulliparous women, and type of previous birth. However, there were statistically significant differences in maternal age, smoking habits during the pregestational period, previous birth weight, and preterm delivery between the two groups (p\u0026thinsp;=\u0026thinsp;0.049, 0.001, 0.001, and 0.001, respectively). Both the maternal age and previous birth weight were lower in pregnant women with SGA, and smoking habits and preterm delivery were also more common in this group. None of the pregnant women included in the study had hypertension, diabetes mellitus, or SLE during the pregestational period, but gestational hypertension and diabetes mellitus were detected in respectively 1 and 6 pregnant women with SGA during follow-up visits. Out of 102 pregnancies, 101 were liveborn babies, and 1 baby born to a pregnant woman with SGA was stillborn.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe demographic and obstetric findings of the patients included in the study.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eParameters\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSGA (+)\u003c/p\u003e \u003cp\u003e(52, 50.9%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSGA (-)\u003c/p\u003e \u003cp\u003e(50, 49.1%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eMaternal age (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26.04\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e4.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27.94\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e5.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.049\u003c/b\u003e\u003csup\u003e\u003cb\u003ea\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePre-pregnancy BMI (kg/m2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29.15\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e2.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30.05\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e4.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.357\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSmoking habits\u003c/p\u003e \u003cp\u003eduring the pre-pregnancy period\u003c/p\u003e \u003cp\u003e(n, %)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (100%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003csup\u003e\u003cb\u003eb\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49 (49.49%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50 (50.51%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eGravida\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.75\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e0.97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.84\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.547\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eParity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.56\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e0.67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.68\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e0.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.650\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003ePrevious type of birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNulliparous\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (50.94%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e26 (49.06%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e0.878\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSVD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (48.48%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17 (51.52%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eC/S\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9 (56.25%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7 (43.75%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePrevious birth weight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2895.40\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e445.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3329,.8\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mp\\:\\)\u003c/span\u003e\u003c/span\u003e323.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003csup\u003e\u003cb\u003ea\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003ePreterm delivery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (78.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (21.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.001\u003c/b\u003e\u003csup\u003e\u003cb\u003eb\u003c/b\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003ea: Mann-Whitney U test (Data is not normal) b: Chi-Square Test\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, maternal weight, length, age, gravida, and parity were identified as feature parameters. After training the data, the test data was used to assess the ML algorithms and it was determined that the LGBM model was the most powerful algorithm with the highest accuracy rate (71.4%) and AUC-ROC (62.7%) for predicting SGA at birth using maternal sociodemographic findings and obstetric history (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePrognosis Prediction Results of Different Machine Learning Algorithms.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"6\" nameend=\"c7\" namest=\"c2\"\u003e \u003cp\u003eModel Training Results\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003eTest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eCross-Validation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel Name\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eROC-AUC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInterval\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eROC-AUC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eInterval\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRandom Forest Classifier\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.591\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.524\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.31\u0026ndash;0.737\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.681\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.605\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.429\u0026ndash;0.781\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAVG Blender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.632\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.642\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.36\u0026ndash;0.783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.584\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.633\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.464\u0026ndash;0.82\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic Regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.593\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.274\u0026ndash;0.682\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.479\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.511\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.412\u0026ndash;0.773\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLGBM Classifier\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.627\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e0.714\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e0.458\u0026ndash;0.847\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e0.589\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e0.568\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003e0.382\u0026ndash;0.754\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExtra Trees Classifier\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.673\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.617\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.192\u0026ndash;0.591\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.666\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.438\u0026ndash;0.797\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eXGB Classifier\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.664\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.465\u0026ndash;0.868\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.637\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.49\u0026ndash;0.843\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eAVG Blender: Average Blender, LGBM: Light Gradient Boosting Machine, XGB: eXtreme Gradient Boosting\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe confusion matrix revealed that 9 out of 11 cases of SGA at birth and 6 out of 10 cases of non-SGA at birth were accurately predicted by the LGBM algorithm in the training set (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe ability to predict intrauterine growth restriction has the potential to change the course of the disease and will help improve monitoring, allow for early identification of the condition, and prevent or reduce fetal growth retardation. Therefore, it is extremely important to identify maternal risk factors that can predict intrauterine growth restriction in the early stages of pregnancy. The purpose of this study was to create a machine-learning model to identify small for gestational age (SGA) by forecasting the probability of pregnant women developing SGA.\u003c/p\u003e \u003cp\u003eRecent studies have suggested that clinical models that include maternal background and Doppler ultrasound in the first or second trimester have shown only moderate success (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). Based on these findings, there is potential for improvement by using artificial intelligence (AI) and machine learning techniques in prenatal risk assessment (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Several studies have been conducted using different artificial intelligence models to detect fetal growth retardation. When additional imaging methods and parameters, such as CTG (cardiotocography), are included in the clinical data, the detection rates for SGA can increase up to 93%. Artificial intelligence and machine learning have demonstrated the potential to enhance predictions, particularly with CTG readings (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). However, there are limitations to the meta-analysis, as not all studies utilized the same input variables and samples. To establish a consensus, clinical prospective studies, such as randomized controlled trials, are necessary. Additionally, conducting more studies that compare AI/ML with traditional diagnostics could aid in enhancing the feature selection process and improving algorithm performance in clinical settings (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eCurrently, artificial intelligence (AI) techniques are being used to improve the accuracy of risk assessment and predict adverse perinatal outcomes. These methods are utilized in various areas, including general pregnancy risk assessment, prenatal diagnosis, pregnancy diagnosis, hypertensive disorders of pregnancy, fetal growth, stillbirth, gestational diabetes, preterm delivery, and route of delivery. A systematic review revealed that artificial neural network (ANN) methods are the most effective AI applications for aiding the prediction of medical conditions (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe application of AI techniques has the potential to assist obstetricians in making informed decisions. These techniques can provide valuable insights and predictive capabilities, which can be used in a number of ways. For example, they can be employed in prenatal screening and diagnosis, route of delivery prediction, postpartum care, and personalized care management. They can also be used to predict preterm delivery. By identifying potential risk factors and using trained and validated software, it is possible to detect complications at an early stage, which could then be prevented. This has been demonstrated in the case of patients with preterm delivery during pregnancy (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe dynamic prediction model discussed in this study can identify the risk factors linked to SGA based on the gestational age of pregnant women. It uses a range of models at various gestational ages for prediction, making it straightforward, quick, and easy to implement. Additionally, it holds significant clinical importance.\u003c/p\u003e \u003cp\u003eThe study is limited by its small sample size, its retrospective nature at a single center, and the fact that it only involved Turkish patients. The study's strength is compromised by the lack of additional imaging methods, the absence of laboratory parameters, and the exclusion of additional calculation methods.\u003c/p\u003e \u003cp\u003eUsing machine learning algorithms, like the LGBM model, shows potential for improving the prediction and treatment of SGA pregnancies. This method is cheaper, simpler, and easier to access than other methods because it does not need extra imaging or laboratory methods and can offer useful information for healthcare providers to intervene sooner and enhance outcomes for mothers and babies. The results of the study emphasize the significance of utilizing AI-driven methods in obstetrics to improve decision-making and patient care in situations of high-risk pregnancy. More research and confirmation of these predictive models could help in creating individualized approaches for identifying and treating SGA pregnancies.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur research shows that AI/ML techniques have the potential to improve the screening for SGA in a more precise and cost-effective way, ultimately enhancing pregnancy outcomes. However, further improvements and refinements to the algorithms are needed before these methods can be used in clinical settings. Additionally, it is crucial to ensure the consistency of diagnostic criteria and quality assessment before widespread adoption.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e Conception and design of study: Z.B., E.A.S., Y.K; analysis of data and writing of the manuscript: Z.B., \u0026Ouml;.\u0026Ccedil;., T.T., A.A.Y.; critical revision of the article for intellectual content: Z.B., E.A.S., Y.K., T.T.\u003c/p\u003e\n\u003cp\u003eAll authors read and approved the final manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEthics approval and consent to participate\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe study was conducted in strict accordance with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Hospital (Ankara Medipol University Ethical Committee 08.01.2024/8). All participants gave written informed consent before enrollment.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis manuscript does not report personal data, such as individual details, so consent for publication is not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests. \u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGordijn, S.J., I.M. Beune, and W. Ganzevoort, \u003cem\u003eBuilding consensus and standards in fetal growth restriction studies.\u003c/em\u003e Best Pract Res Clin Obstet Gynaecol, 2018. \u003cstrong\u003e49\u003c/strong\u003e: p. 117-126.\u003c/li\u003e\n\u003cli\u003eMelamed, N., et al., \u003cem\u003eFIGO (International Federation of Gynecology and Obstetrics) initiative on fetal growth: Best practice advice for screening, diagnosis, and management of fetal growth restriction.\u003c/em\u003e International Journal of Gynaecology and Obstetrics, 2021/03. \u003cstrong\u003e152\u003c/strong\u003e(Suppl 1).\u003c/li\u003e\n\u003cli\u003eJaddoe, V.W., et al., \u003cem\u003eFirst trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study.\u003c/em\u003e BMJ, 2014. \u003cstrong\u003e348\u003c/strong\u003e: p. g14.\u003c/li\u003e\n\u003cli\u003eCrovetto, F., et al., \u003cem\u003eFirst-trimester screening for early and late small-for-gestational-age neonates using maternal serum biochemistry, blood pressure and uterine artery Doppler.\u003c/em\u003e Ultrasound Obstet Gynecol, 2014. \u003cstrong\u003e43\u003c/strong\u003e(1): p. 34-40.\u003c/li\u003e\n\u003cli\u003eGordijn, S.J., et al., \u003cem\u003eConsensus definition of fetal growth restriction: a Delphi procedure.\u003c/em\u003e Ultrasound Obstet Gynecol, 2016. \u003cstrong\u003e48\u003c/strong\u003e(3): p. 333-9.\u003c/li\u003e\n\u003cli\u003eAnon, \u003cem\u003eACOG Practice Bulletin No. 204: Fetal Growth Restriction.\u003c/em\u003e Obstetrics \u0026amp; Gynecology, 2019. \u003cstrong\u003e133\u003c/strong\u003e(2): p. e97-e109.\u003c/li\u003e\n\u003cli\u003eLees, C.C., et al., \u003cem\u003eISUOG Practice Guidelines: diagnosis and management of small-for-gestational-age fetus and fetal growth restriction.\u003c/em\u003e Ultrasound Obstet Gynecol, 2020. \u003cstrong\u003e56\u003c/strong\u003e(2): p. 298-312.\u003c/li\u003e\n\u003cli\u003eSalomon, L.J., et al., \u003cem\u003eISUOG Practice Guidelines: ultrasound assessment of fetal biometry and growth.\u003c/em\u003e Ultrasound Obstet Gynecol, 2019. \u003cstrong\u003e53\u003c/strong\u003e(6): p. 715-723.\u003c/li\u003e\n\u003cli\u003eZheng, C., et al., \u003cem\u003eConstruction of prediction model for fetal growth restriction during first trimester in an Asian population.\u003c/em\u003e Ultrasound in Obstetrics \u0026amp; Gynecology, 2024. \u003cstrong\u003e63\u003c/strong\u003e(3): p. 321-330.\u003c/li\u003e\n\u003cli\u003eAkazawa, M., et al., \u003cem\u003eMachine learning approach for the prediction of postpartum hemorrhage in vaginal birth.\u003c/em\u003e Sci Rep, 2021. \u003cstrong\u003e11\u003c/strong\u003e(1): p. 22620.\u003c/li\u003e\n\u003cli\u003eCordina, M., et al., \u003cem\u003eMaternal hemoglobin at 27-29 weeks\u0026apos; gestation and severity of pre-eclampsia.\u003c/em\u003e J Matern Fetal Neonatal Med, 2015. \u003cstrong\u003e28\u003c/strong\u003e(13): p. 1575-80.\u003c/li\u003e\n\u003cli\u003eJavaid, M., et al., \u003cem\u003eSignificance of machine learning in healthcare: Features, pillars and applications.\u003c/em\u003e International Journal of Intelligent Networks, 2022. \u003cstrong\u003e3\u003c/strong\u003e: p. 58-73.\u003c/li\u003e\n\u003cli\u003eRescinito, R., et al., \u003cem\u003ePrediction Models for Intrauterine Growth Restriction Using Artificial Intelligence and Machine Learning: A Systematic Review and Meta-Analysis.\u003c/em\u003e Healthcare (Basel), 2023. \u003cstrong\u003e11\u003c/strong\u003e(11).\u003c/li\u003e\n\u003cli\u003eLian, C., et al., \u003cem\u003eDynamic prediction model of fetal growth restriction based on support vector machine and logistic regression algorithm.\u003c/em\u003e Front Surg, 2022. \u003cstrong\u003e9\u003c/strong\u003e: p. 951908.\u003c/li\u003e\n\u003cli\u003eWarmerdam, G.J.J., et al., \u003cem\u003eDetection rate of fetal distress using contraction-dependent fetal heart rate variability analysis.\u003c/em\u003e Physiol Meas, 2018. \u003cstrong\u003e39\u003c/strong\u003e(2): p. 025008.\u003c/li\u003e\n\u003cli\u003eFeduniw, S., et al., \u003cem\u003eApplication of Artificial Intelligence in Screening for Adverse Perinatal Outcomes-A Systematic Review.\u003c/em\u003e Healthcare (Basel), 2022. \u003cstrong\u003e10\u003c/strong\u003e(11).\u003c/li\u003e\n\u003cli\u003eRittenhouse, K.J., et al., \u003cem\u003eImproving preterm newborn identification in low-resource settings with machine learning.\u003c/em\u003e PLoS One, 2019. \u003cstrong\u003e14\u003c/strong\u003e(2): p. e0198919.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Small for gestational age (SGA), artificial intelligence, machine learning, prediction, LGBM model, obstetrics, high-risk pregnancy","lastPublishedDoi":"10.21203/rs.3.rs-4850407/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4850407/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSmall for gestational age (SGA) is a significant concern in obstetrics, with implications for stillbirth, neonatal mortality, and long-term health outcomes. Early detection of SGA is crucial for prevention and treatment, but current methods have limitations. This study aimed to develop an artificial intelligence (AI)-based algorithm to predict SGA using sociodemographic and obstetric features during pregnancy.\u003c/p\u003e \u003cp\u003eA total of 102 pregnant women meeting specific criteria were included in the study. The feature impact factors considered important factors for predicting SGA at birth were maternal weight, length, age, gravida, and parity. The LGBM model demonstrated the highest accuracy rate (71.4%) and AUC-ROC (62.7%) in predicting SGA, showcasing its potential for improving the prediction and treatment of SGA pregnancies.\u003c/p\u003e \u003cp\u003eThe study highlights the importance of using AI-driven methods in obstetrics to improve decision-making and patient care in high-risk pregnancy scenarios. Although AI/ML techniques have shown promise in enhancing the screening for SGA, further refinement and validation of algorithms are necessary before clinical implementation. Consistency in diagnostic criteria and quality assessment is essential for the widespread adoption of these methods in clinical settings.\u003c/p\u003e","manuscriptTitle":"Role of artificial intelligence in the prediction of small-for-gestational-age birth weight by gestational week","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-09 09:41:31","doi":"10.21203/rs.3.rs-4850407/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"09ad97df-2a57-4d0b-8c94-a80c0f3f6a5e","owner":[],"postedDate":"September 9th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-12-30T23:53:13+00:00","versionOfRecord":[],"versionCreatedAt":"2024-09-09 09:41:31","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4850407","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4850407","identity":"rs-4850407","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}