Precision Timing of Delivery in Fetal Growth Restriction: A Systematic Review of Integrated Biomarker-Imaging Prognostic Models and Decision Algorithms to Balance Stillbirth and Prematurity Risk

ABSTRACT The timing of delivery in pregnancies complicated by fetal growth restriction (FGR) remains one of the most challenging decisions in perinatal medicine. Clinicians must navigate a narrow margin between the rising threat of stillbirth and the risks associated with preterm birth, yet current guidance often relies on fragmented evidence or single-modality assessments. To clarify the state of knowledge and identify pathways toward precision-based care, we conducted a systematic review integrating data from angiogenic biomarkers, Doppler velocimetry, fetal monitoring technologies, growth-trajectory analytics, and emerging machine-learning approaches. Following PRISMA 2020 guidance, we searched major databases, registers, and reference lists, screened 2,480 records, and included 40 studies that met predefined criteria related to diagnostic performance, prognostic modeling, or delivery-timing algorithms in FGR. Across the literature, several converging patterns emerged. Low placental growth factor and elevated sFlt-1 consistently anticipated Doppler deterioration in early-onset FGR, while umbilical artery, middle cerebral artery, and ductus venosus waveforms provided sequential hemodynamic cues that aligned closely with adverse outcome risk. Late-onset FGR showed a different trajectory, with more subtle biomarker deviations and greater reliance on cerebral redistribution and growth-velocity decline. Incorporation of machine learning–particularly models combining angiogenic markers with Doppler indices–demonstrated improved short-term risk prediction, though external validation remains limited. Overall, the evidence supports a shift toward integrated biomarker-imaging prognostication to guide delivery timing. However, heterogeneity in definitions, surveillance thresholds, and analytic methods underscores the need for harmonized frameworks, prospective validation, and adaptive decision algorithms capable of balancing fetal and neonatal risks in real time.