A Review on the Use of Artificial Intelligence Algorithms for the Prediction of Fetal Brain and Heart Abnormalities

Abstract

Artificial intelligence is an essential tool in the fight against congenital foetal anomalies in the field of foetal medicine. With the 
use of ML algorithms and CNNs, abnormalities of the heart and brain in fetal magnetic resonance imaging (MRI) and ultra
sonography may be identified, detected, and localized. Prenatal abnormalities may be better identified and predicted with the 
use of Artificial Intelligence (AI) systems that can do complex assessments of abnormal picture patterns. This narrative review 
delves into the use of AI in congenital anomaly detection and risk stratification. Machine learning and deep learning algorithms 
have the potential to enhance fetal imaging (ultrasonography and MRI) exams in order to shorten examination times, lessen 
the burden on the clinician, and improve the accuracy of diagnoses for fetal abnormalities. The purpose of this research is to 
assess the efficacy of the algorithms already in use to automate the detection of abnormalities in the developing brain and heart 
of pregnant women. The article also contrasts ML and DL algorithms by looking at how well they identify abnormalities in the 
developing foetus’s brain and heart. For the purpose of congenital anomaly prediction, the study stresses the need for bigger 
and more diverse datasets, analysis of longitudinal data, and integration of several data sources. Additional emphasis is placed 
on the value of interpretability, human clinical experience, and prospective validation in actual clinical contexts