Role of Inflammatory Markers and Doppler Parameters in Late-Onset Fetal Growth Restriction: A Machine Learning Approach

Overview

Fetal growth restriction (FGR) is a serious complication in pregnancy that can lead to various adverse outcomes. It's classified into early-onset (before 32 weeks) and late-onset (after 32 weeks), with late-onset associated with long-term risks like hypoxemia and developmental delays. The study focuses on the role of inflammation in FGR, introducing new blood markers for better understanding and diagnosis. It also addresses the challenges of using advanced diagnostic tools in low-resource settings and explores the use of machine learning to predict FGR based on inflammatory markers, highlighting the potential of artificial intelligence in overcoming these challenges.

Fetal growth restriction (FGR), also known as intrauterine growth restriction, is a prevalent pregnancy complication with potentially negative outcomes for newborns. The condition's causes are varied, involving genetic factors, maternal inflammation, infections, and other pathologies. FGR is categorized based on its onset: early-onset FGR occurs before 32 weeks' gestation, while late-onset happens after 32 weeks. Late-onset FGR, though less risky in perinatal complications compared to early-onset, is linked to an increased risk of hypoxemia and neurodevelopmental delays. Diagnosis primarily relies on ultrasound measurements and Doppler flow analysis of specific arteries. The study highlights the complexity of diagnosing and managing late-onset FGR, emphasizing the unclear pathophysiological mechanisms. It proposes the exploration of inflammatory processes and the potential role of new markers such as the systemic immune inflammation index (SII), systemic inflammatory response index (SIRI), and neutrophil-percentage-to-albumin ratio (NPAR) for understanding FGR. These markers are easily measured through blood tests and are significant in various diseases. The text also discusses the challenges of applying advanced diagnostic methods in low-income countries due to the need for sophisticated equipment, contrasting with the accessibility of artificial intelligence and machine learning models via the internet. The study aimed to assess the impact of inflammatory processes on late-onset FGR by analyzing NPAR, along with other markers, and evaluating their predictive value using machine learning algorithms.