EVALUATING THE ACCURACY AND CLINICAL UTILITY OF AI ALGORITHMS FOR GESTATIONAL AGE PREDICTION: A COMPREHENSIVE SYSTEMATIC REVIEW

Abstract

Background: Accurate gestational age (GA) estimation is vital for optimal prenatal care.  This systematic review evaluates the accuracy and clinical utility of artificial intelligence (AI) algorithms for GA prediction using ultrasound data.  By synthesizing evidence from diverse populations and imaging approaches, this review highlights current performance, potential benefits, and limitations, guiding future research and clinical adoption. Methods: This systematic review was conducted in accordance with PRISMA guidelines.  A comprehensive literature search was performed across major databases, including PubMed, Scopus, Web of Science, and Google Scholar, to identify relevant studies on AI-based gestational age prediction. Results: This systematic review identified 22,350 records, including 22,300 from database searches and 50 from other sources. After removing duplicates, 18,400 records were screened, and 17,900 were excluded based on titles and abstracts.  Of the 500 full-text articles assessed, 492 were excluded for not meeting inclusion criteria.  Finally, 8 studies were included in the qualitative synthesis, providing valuable evidence on the accuracy and clinical utility of AI algorithms for gestational age prediction. Conclusion: . This review demonstrates that AI algorithms have the potential to improve prenatal care, particularly in settings with limited resources, and that their accuracy for predicting gestational age is encouraging. However, in order to support widespread adoption and confirm their clinical utility, additional large-scale studies and external validations are required.

Keywords: Artificial intelligence, Machine learning,  Deep learning, Gestational age prediction, Neural networks, convolutional neural networks (CNN) ,Support vector machines (SVM), Algorithm validation