(211) Artificial Intelligence-driven precision pathology identifies distinct placental findings in severe fetal growth restriction or hypertension

To leverage artificial intelligence (AI)-driven precision pathology to quantify the effects of severe FGR, pregnancy-related HTN, or their interaction on placental villus and vascular properties. We hypothesize that while only villi are affected in HTN, both villous and vascular structures are diminished in severe FGR with no additional effect of concomitant HTN.

Study Design:

In this retrospective cohort study, paraffin-embedded placental sections were obtained from four groups: [1] Severe FGR with HTN (n=11) [2] Severe FGR with normotension (n=14) [3] Appropriately grown, gestational age (GA)-matched pregnancies with HTN (n=12) and [4] Appropriately grown, GA-matched pregnancies with normotension (n=14). These underwent immunodetection for cytokeratin-7 (trophoblast) and CD34 (endothelium) followed by AI-driven image analyses. Number of villi and placental vessels, total villous area and total vascular area were quantified. Outcomes were normally distributed, and quantitative analyses were modeled with linear regression models.

Results:

Pregnancies with HTN exhibited significantly less stem villi (p < 0.01), stem villous area (p < 0.01), and stem villous vessels (p < 0.01), with no difference in vascular area. In contrast, abnormalities in severe FGR were limited to terminal villi, with fewer villi and vessels (p < 0.01) and decreased villous area (p < 0.01) and vascular area (p=0.02). No significant effects were found with the interaction of HTN and severe FGR.

Conclusion:

Pregnancy-related HTN and severe FGR are associated with stem and terminal villus anomalies, respectively, suggesting that perturbations in placental development may be occurring at different gestational periods for each condition.