Preeclampsia (PE) is a common pregnancy complication affecting 3%–8% of pregnancies globally [1, 2]. It remains a leading cause of maternal and perinatal morbidity and mortality [1, 2]. Worldwide, approximately 10 million women develop PE annually, resulting in an estimated 76,000 maternal deaths and 500,000 neonatal deaths attributable to PE-related hypertensive disorders [1, 3]. According to the American College of Obstetricians and Gynecologists, the principal diagnosis of PE is the onset of hypertension after 20 weeks of gestation, accompanied by signs of end-organ damage including proteinuria, thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or new-onset headache/visual symptoms that are unresponsive to medication and unexplained by alternative diagnoses [4]. Clinically, PE is stratified into early-onset (34 weeks of gestation) and late-onset ($\ge$34 weeks of gestation) subtypes based on the gestational age (GA) at disease onset [5].

Placental perfusion is associated with both maternal blood flow and fetal circulation. Physiologically, trophoblast invasion induces remodeling of maternal uterine spiral arteries, characterized by loss of the vascular muscular layer and subsequent vasodilation. This transformation establishes high-flow uteroplacental circulation through the structural conversion of uterine arteries into low-resistance vessels [1, 6, 7]. The pathophysiology of early-onset PE stems primarily from defective spiral artery remodeling, leading to impaired uteroplacental perfusion, defective placentation, and subsequent placental hypoxia [8, 9]. In contrast, late-onset PE is associated with placental oxidative changes triggered by progressive disparity between maternal perfusion and fetoplacental demands [8, 9]. While the pathogenesis and placental pathology of early- and late-onset PE appear to be different, both forms involve placental dysfunction and ischemia [9, 10].

Given the central role of the placenta in PE pathophysiology, it is important to develop non-invasive, clinically practical methods for its in vivo functional assessment [11, 12]. In terms of conventional ultrasonography, Doppler parameters for the umbilical artery (UA) and uterine artery (UtA) provide overall resistive indices for the upstream arterial vessels of placental vascular beds on the fetal and maternal sides, respectively. However, these metrics only provide an indirect assessment of placental perfusion and cannot directly quantify microvascular density within the placenta [13]. Therefore, more advanced non-invasive imaging modalities capable of direct visualization and quantitative assessment of placental microvascular perfusion characteristics may be required [14].

Previous studies have attempted to assess placental perfusion using three-dimensional (3D) power Doppler ultrasound. However, this approach is limited by factors such as equipment settings, the requirement for experienced operators, and low reproducibility in the second and third trimesters [13, 14]. Recently, a new and non-invasive microvascular-flow (MV-Flow) imaging technology (Samsung Medison Co. Ltd, Seoul, Korea) has been used to assess the microvasculature in carotid plaque, hepatic parenchyma, fetal brain, and the placenta in normal, fetal growth restriction, and placenta accreta spectrum disorders [13]. Thanks to key features including high tissue suppression (noise reduction), advanced flash artifact filtering, compound imaging capabilities, and high sensitivity, MV-Flow technology can detect minute vessels with low-velocity blood flow that are not visible with conventional Doppler imaging [15, 16]. This innovative technology not only enables visualization of the placental villous shaft, but also quantification of the placental microvascular structure at the level of the villous lobe [17].

The aim of the present study was to evaluate the utility of MV-Flow imaging in characterizing the placental microvascular architecture in PE pregnancies. In addition, considering the distinct pathophysiological pathways between early- and late-onset PE, we compared microvascular differences between PE and normal pregnancies, as well as between early- and late-onset PE cases.

Among the 87 cases included in this study, 50 were normal pregnancies and 37 were diagnosed with PE. Women with PE were significantly older than healthy controls (31.40 $\pm$ 5.50 vs. 28.30 $\pm$ 3.90 years; p = 0.005; Table 1). At the time of ultrasound examination, the GA of the normal group (30.80 $\pm$ 4.30 weeks) was not significantly different to that of the PE group (32.20 $\pm$ 3.40 weeks; p = 0.102; Table 1).

The intra-observer correlation coefficients for VI^MV-m and VI^MV-p were 0.96 (95% confidence interval (CI): 0.94–0.97) and 0.94 (95% CI: 0.92–0.96), respectively, demonstrating the high reliability of measurements.

In normal pregnancies, no significant difference was observed between the VI^MV-p and VI^MV-m (42.80 $\pm$ 8.50 vs. 42.40 $\pm$ 8.10, p = 0.556). Moreover, neither the VI^MV-p nor the VI^MV-m was correlated with GA (p = 0.410 and 0.086, respectively). Similarly, in PE pregnancies, no significant difference was observed between the VI^MV-p (19.90 [18.40–24.70]) and VI^MV-m (21.20 [17.30–27.40], p = 0.829), and neither parameter was correlated with GA (p = 0.455 and 0.207, respectively).

VI^MV values for the control and PE groups were compared irrespective of the GA. Both the VI^MV-p and VI^MV-m in the PE group were significantly lower than those in the control group (p 0.001; Table 1).

GA at examination was 34 weeks in 63 cases (38 normal and 25 early-onset PE cases), and $\ge$34 weeks in 24 cases (12 normal and 12 late-onset PE cases). For cases 34 weeks, no significant difference in GA was observed between the normal and early-onset PE groups (p = 0.115; Table 2). However, the early-onset PE group exhibited significantly lower VI^MV-p and VI^MV-m values compared to normal controls (p 0.001; Table 2). For cases $\ge$34 weeks, no significant difference in GA was observed between the normal and late-onset PE groups (p = 0.779; Table 3). However, the late-onset PE group also showed significantly lower VI^MV-p and VI^MV-m values compared to the normal group (p 0.001; Table 3; Fig. 2). No significant differences in placental VI^MV-p (p = 0.170) or VI^MV-m (p = 0.471) were found between early- and late-onset PE cases.

Fig. 2.

A late-onset PE case at 37⁺⁰ weeks of gestation. (A) VI^MV for the middle segment (VI^MV-m, 27.8%). (B) VI^MV for the peripheral segment (VI^MV-p, 22.8%). The yellow circle shows the ellipse-shaped region of interest for which the VI^MV was calculated.

Hemodynamic parameters for the UA, bilateral UtAs, and MCA were compared between PE cases and GA-matched normal cases. In the early-onset PE group, the RI and PI of the UA and of both UtAs were significantly higher than those of the normal group at 34 weeks GA (p 0.05; Table 2). In contrast, the MCA-RI was significantly lower (p 0.05; Table 2). In the late-onset PE group, UA-RI and UA-PI were elevated compared to normal cases $\ge$34 GA (p 0.05; Table 3), but no significant differences were observed between the two groups for the UtA or MCA hemodynamic parameters (p $>$ 0.05; Table 3). The centiles of UA-PI and UtA-PI according to GA were further evaluated against recently published reference ranges (Tables 2,3) [22, 23].

Perinatal outcomes were systematically evaluated in pregnancies complicated by PE. In the 37 PE cases, two early-onset cases were lost to follow-up, while three early-onset cases underwent medically indicated pregnancy termination. All 32 remaining PE cases (20 early-onset and 12 late-onset) underwent cesarean delivery at our institution, allowing comprehensive analysis of perinatal outcomes including GA at delivery, birth weight, and Apgar scores (1-, 5-, and 10-minute) (Table 4). The Apgar scores at 1 and 5 minutes were significantly higher in the late-onset PE group than in the early-onset PE group (p 0.001 and p = 0.039, respectively). Based on birth weight and GA at delivery, 24 neonates were classified as AGA and 8 as SGA. No significant differences in placental vascular indices were observed between the AGA and SGA groups: VI^MV-p (21.40 [17.70–27.30] vs. 21.50 $\pm$ 2.50, p = 0.949), and VI^MV-m (21.30 [17.50–27.50] vs. 22.80 $\pm$ 6.40, p = 0.881).

This study utilized MV-Flow ultrasound technology to quantify placental microvascularization, expressed as VI^MV, in both normal and PE pregnancies. Our findings revealed significantly reduced placental VI^MV values in PE pregnancies compared to normal pregnancies. Compared with normal pregnancies of 34 weeks GA, early-onset PE cases exhibited significantly lower placental VI^MV, as well as greater hemodynamic resistance in both UtAs and UA. Compared with normal pregnancies of $\ge$34 weeks GA, late-onset PE cases demonstrated lower VI^MV and higher UA resistance; however, no significant difference was observed in the resistance of UtAs between these groups.

We found no difference between the VI^MV-p and VI^MV-m in either the PE or normal groups, demonstrating that placental VI^MV was not related to the measurement site. Both VI^MV-m and VI^MV-p showed high intra-observer agreement, with correlation coefficients of 0.96 and 0.94, respectively. These results indicate that placental VI^MV could be a highly reproducible metric, regardless of the sampling location. Chen et al. [13] also previously evaluated the placental microvascular architecture in normal and fetal growth-restricted pregnancy with MV-Flow imaging. In agreement with the present study, these workers found no significant variation in VI^MV among different placental regions.

Our analysis demonstrated that placental VI^MV was not associated with GA in either normal pregnancies or PE cases. Previous studies employing 3D power Doppler with virtual organ computer-aided analysis (VOCAL) technique demonstrated that placental vascularization indices (Vascularization Index, Flow Index, Vascularization Flow Index) were not dependent on GA during mid- and late-trimester [24, 25]. Our quantitative analysis revealed a significant reduction in placental VI^MV among PE pregnancies compared to normal pregnancies. Previous studies that utilized placental 3D power doppler also showed decreased placental vascularization indices in pregnancies with PE [11, 26]. Therefore, despite using different ultrasound techniques, the present study and others have consistently shown that placental vascularization does not change according to the GA. Furthermore, these studies have consistently demonstrated impaired placental vascularization in PE pregnancies, concordant with the pathological feature of placental ischemia in PE [10, 11, 26].

In early-onset PE cases, we also found that both the RI and PI of the UA and UtAs were significantly higher compared to normal pregnancies at GA 34 weeks. In contrast, no significant differences were observed in UtA hemodynamics between late-onset PE cases and GA-matched controls ($\ge$34 weeks), although late-onset PE exhibited elevated RI and PI of the UA. These divergent hemodynamic characteristics between early- and late-onset PE subtypes may be related to different pathogenic mechanisms. Early-onset PE is probably caused by insufficient placentation secondary to failed spiral artery remodeling, manifesting as elevated UtA resistance, whereas late-onset PE is more likely due to an imbalance between decreased placental function and maternal perfusion and fetoplacental demands [8, 9].

As previously demonstrated, both early- and late-onset PE cases exhibited significantly higher UA resistance indices compared to GA-matched controls, along with reduced placental microvascular density as evidenced by markedly lower placental VI^MV. Together, these findings confirm the pathological microvascular insufficiency characteristic of PE placenta. Notably, UA resistance serves as an indirect indicator of placental hypoperfusion, whereas VI^MV provides direct quantitative evidence of microvascular deficiency.

However, our study found no significant differences in placental VI^MV-p (p = 0.170) or VI^MV-m (p = 0.471) between early- and late-onset PE cases. This indicates that although the early- and late-onset subtypes have different pathophysiological pathways, both subtypes may have similar impairment of placental perfusion.

The perinatal outcomes of PE pregnancies were also analyzed in this study. In PE cases, no significant differences in placental VI^MV were observed between AGA and SGA neonates, suggesting that VI^MV may lack predictive value for SGA development in PE pregnancies. Using alternative placental vascular imaging techniques, the existing literature presents conflicting evidence regarding vascular indices and SGA prediction. Some studies reported significant associations between reduced placental vascularization and SGA risk [27, 28]. However, consistent with our findings, other investigators [29, 30] found that placental vascular indices cannot predict this complication. Therefore, it remains controversial as to whether reduced placental vascularization can predict the risk of SGA newborns.

The primary novelty of our study lies in its application for the first time of MV-Flow imaging to quantitatively evaluate and compare the placental microvasculature between normal and PE pregnancies, and between early- and late-onset PE cases. Other recent studies have employed the microvascular imaging technique of Superb Microvascular Imaging (SMI) to evaluate the placental vascular characteristics [14, 20, 31, 32, 33, 34]. These studies were mainly conducted to evaluate placental microvascularization in normal pregnancy, pregnancy with fetal growth restriction, pregnancy with gestational diabetes, and placenta accreta spectrum [14, 20, 31, 32, 33, 34]. SMI also proved useful for the evaluation of placental microcirculation. However, those SMI studies did not specifically examine differences in the placental microvasculature between normal and PE pregnancy, or between early- and late-onset PE.

MV-Flow imaging enables quantitative assessment of the microvascular architecture in PE placentas. Both early- and late-onset PE pregnancies showed significantly lower placental VI^MV compared to normal pregnancies, thus supporting the clinical utility of MV-Flow for evaluation of placental perfusion in PE. Furthermore, no significant difference in placental VI^MV was observed between early- and late-onset PE, suggesting similar impairment of placental perfusion in both subtypes.

The data that support the findings of this study are available on request from the corresponding author.

XZ, YF, and LL designed the research study. JP, YG, and XZ performed the research. LM, MH, and JS analyzed the data. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was carried out in accordance with the guidelines of the Declaration of Helsinki. The study was approved by the Research Ethics Committee of Shaanxi Provincial People’s Hospital (2021213) and written informed consent was obtained from all subjects.

This work was supported by the Science and Technology Talents Support Project of Shaanxi Provincial People’s Hospital (No. 2022JY-22), the Social Development Projects of Science and Technology Department of Shaanxi Province (No. 2022SF-209), the Project of Shaanxi Provincial People’s Hospital (No. 2023YJY-62), and the Natural Science Basic Research Program of Science and Technology Department of Shaanxi Province (No. 2025JC-YBQN-1234).

The authors declare no conflict of interest.