Integrated quantitative biomarkers of placental texture and hemodynamics improve assessment of pregnancies at risk for adverse fetal outcomes.

INTRODUCTION: Globally, more than two million babies are stillborn every year. The primary cause of stillbirth is placental dysfunction, which impairs oxygen and nutrient delivery to the fetus. From a physical perspective, the efficiency of this exchange depends on placental blood flow and microvascular villous structure. Doppler ultrasound (US) imaging is used clinically to assess abnormal placental blood flow, but evaluation of placental structure remains largely observational. Texture analysis of B-mode US images offers a quantitative approach to indirectly characterize placental structure through analysis of US image pixels. Therefore, the objective of this study was to investigate whether integrated quantitative biomarkers derived from placental texture, Doppler ultrasound, and standard clinical variables improve the assessment of pregnancies associated with adverse fetal outcomes. METHODS: Patients (N = 129) were recruited, and B-mode US images and Doppler measurements were acquired at visit 1 (28-32 weeks of gestation) and visit 2 (34-38 weeks of gestation). Clinical outcome data were used to define patients as High-Risk based on previously established markers of significant stillbirth risk (e.g., preeclampsia with severe features, severe FGR, perinatal hypoxia). Logistic regression and random forest models were used to evaluate the predictive value of each modality with clinical variables and in combination. RESULTS: Quantitative placental texture features demonstrated predictive performance comparable to Doppler ultrasound measurements, supporting their associations with adverse fetal outcomes. Integrating clinical, Doppler, and texture features further improved model performance compared with models developed from clinical variables alone or paired with either imaging modality individually. These findings demonstrate that quantifying placental US texture provides complementary information to Doppler and clinical variables. While further validation in larger external cohorts is required, this work supports the potential of quantitative multimodal ultrasound biomarkers for assessing pregnancies with adverse fetal outcomes.