Introduction

Preeclampsia (PE) is a multisystem disorder of pregnancy affecting 3–5% of pregnancies [1]. It is a major cause of perinatal morbidity and mortality [2]. PE is defined by the International Society for the Study of Hypertension in Pregnancy (ISSHP) as the occurrence of hypertension that develops after the 20th week of pregnancy, accompanied by proteinuria or signs of acute kidney injury, liver dysfunction, neurological manifestations, hemolysis, thrombocytopenia, or fetal growth restriction (FGR) [3]. The maternal vascular malperfusion in PE is characterized histologically by diminished placental size, infarction, abnormal development of the placental villi, and a deficiency of the maternal decidual spiral arterioles [4, 5]. Although the cause of PE is unknown, the accumulated evidence strongly implicates the placenta [6, 7].

As PE is closely related to placental dysfunction, the antenatal assessment of placental function is crucial to understand this complex condition [8]. Placental imaging offers a window into the placental contribution and mechanisms. Studies have reported that the mean placental stiffness detected using ultrasound is considerably higher in preeclamptic pregnancies than in controls in the third trimester [9]. The elasticity of the in vivo placentas used to be detected by ultrasound shear wave elastography (SWE) because magnetic resonance elastography is not recommended for pregnant patients. However, the accuracy of SWE measurement depends on the technique and experience of the operator. Moreover, the posterior placenta is hard to obtain adequate measurements.

In this study, we employed virtual magnetic resonance elastography (vMRE) for assessing placental stiffness. The vMRE used a diffusion-weighted imaging (DWI) sequence to enhance the texture of tissues with dense diffusion barriers, such as fibrotic tissues. Le Bihan et al. proposed a DWI-based vMRE to evaluate liver fibrosis. They demonstrated a remarkable correlation between the DWI-based shear modulus (μdiff) and the MRE shear modulus (μMRE) [10]. A few studies have suggested that vMRE can be used to measure the stiffness of soft tissues [11,12,13,14,15,16]. In recent years, this approach has also been used in studies on the placenta [17]. However, the value of vMRE for detecting the placental stiffness of pregnancies with PE is still unknown. Therefore, this study aimed to explore the value of DWI-based elasticity in pregnancies with PE.

Materials and methods

Study population

This prospective case–control study was approved by the institutional ethics committee in our hospital (Ethics approval number: KS23253). Seventy-four PE pregnancies and eighty-two control pregnancies participated in this study were recruited from January 2023 to September 2023. We successfully obtained the informed written consent from each participant. The inclusion criteria were as follows: (1) the maternal age ≥ 20 years old; (2) singleton pregnancies of 20 weeks gestation or later, (3) Preeclampsia was prospectively defined using the international consensus definition: gestational hypertension accompanied by one or more of the following new-onset conditions at or after 20 weeks’ gestation: proteinuria, acute kidney injury, liver involvement, neurological complications, hematological complications, and uteroplacental dysfunction [18]. The normal control group of pregnancies consisted of healthy women with normal ultrasound results and birthweight. They underwent MRI during 21–35 weeks’ gestation.

The exclusion criteria were as follows: (1) poor-quality MRI images with severe fetal movement; (2) chronic hypertension; (3) placental abnormalities including placental implantation, placental previa, and placental abruption; (4) metallic implants; and (5) claustrophobia. The control group of pregnant women when the following criteria were fulfilled: singleton pregnancies of 20 weeks gestation or later, no diagnosis of a hypertensive disorder at enrollment and until delivery, no significant past medical history, no pregnancy complications (including gestational diabetes mellitus), delivery at term with birth weight between the 3rd and 97th centiles (calculated using International Fetal and Newborn Growth Consortium for the 21st Century version 1.3.5). A total of 60 pregnant women were enrolled in the final analysis, including 35 controls and 25 pregnant women with PE (Fig. 1). vMRE measurements were obtained upon diagnosis of PE in the outpatient or inpatient setting. All patients received routine care for managing PE, which was based on established protocols and the clinical expertise of the supervising obstetrician. This included expectant management, treatment of hypertensive emergencies with anti-hypertensive medications, and use of magnesium sulfate for seizure prophylaxis or delivery when deemed necessary.

Fig. 1
figure 1

Flow diagram of patient recruitment of the study

Full size image

Magnetic resonance imaging protocol

All patients were examined on a 1.5-T scanner (MAGNETOM Aera, Siemens Healthineers AG, Erlangen, Germany). The magnetic resonance imaging (MRI) protocols were as follows: T1-weighted VIBE sequence [repetition time (TR)/echo time (TE) = 4.85 ms/2.38 ms slice thickness = 5.0 mm; field of view (FOV) = 400 × 368 mm2; flip angle (FA) = 10°; and in-plane resolution = 1.5 × 1.5 mm2]; T2-weighted half-Fourier-acquired single-shot turbo spin echo sequence (TR/TE = 1300 ms/109 ms; slice thickness = 5.0 mm; FOV = 400 × 309 mm2; FA = 120°, and in-plane resolution = 1.5 × 1.5 mm2); and multi-b value DWI sequence (TR/TE = 6400 ms/65 ms; slice thickness = 5 mm; FOV = 320 × 320 mm2; and in-plane resolution = 3.5 × 3.5 mm2) with a spectrum of varying b values of 50, 200, and 800 s/mm2. The total scan duration was less than 10 min.

Measurement of placental stiffness

The region of interest (ROI) segmentation was performed using the open-source software ITK-SNAP [18]. The stiffness value of DWI-based vMRE was determined using custom-written software in MATLAB (MathWorks, MA, USA). DWI of the lower b value (Slow, b value = 200 s/mm2) and that of the higher b value (Shigh, b value = 800 s/mm2) were used to estimate the virtual stiffness presented by the μdiff: µdiff = α·(ln Slow/Shigh) + β

where α is the scaling factor and β is the shift factor, set to − 9.8 and 14, respectively, according to the previous calibration studies on the liver [19]. The mean values of placental μdiff were automatically extracted from the segmented placenta regions using the custom-written software.

Statistical analysis

The statistical analyses were performed using SPSS software (version 27.0). The Student t test was used to perform pairwise comparisons. The χ2 test was used to test the association between categorical variables. Multiple regression analysis was used to assess the relationship between placental stiffness and PE. The receiver-operating characteristic (ROC) curve analysis and the area under the curve (AUC) were used to quantify and compare the diagnostic value of each remarkable parameter. The potential relationship between μdiff and gestational week at MRI in normal pregnancies was explored using a linear regression model. A P value < 0.05 indicated a statistically significant difference.

Results

Baseline characteristics

The details regarding the baseline and clinical characteristics of the participants are summarized in Table 1. No substantial differences in the mean maternal age, body mass index (BMI) at admission, or mean gestational age (GA) at the time of MRI scanning were noted between the control and PE groups. The PE group was associated with a significantly increased rate of preterm delivery (52 vs 14.3%; P = 0.009) and reduced final birthweight (2138.57 ± 837 g vs 3299.10 ± 539 g; P = 0.014) (Table 1).

Table 1 Baseline of the control group (n = 35) and the PE group (n = 25)
Full size table

Comparison of ADC and μ diff between control and PE groups

The mean ADC value was lower in the PE group compared with the control group (1.414 ± 0.228 × 10–3 mm2/s vs 1.737 ± 0.107 × 10–3 mm2/s, P = 0.031) (Fig. 2; Table 2). The mean μdiff value was 5.901 ± 1.757 kPa and 4.618 ± 2.055 kPa for the PE and control groups, respectively (Fig. 2; Table 2). The AUC of the ROC curve was 0.903 and 0.796 for the PE and control groups, respectively (Fig. 3; Table 2). The optimized cut-off value of placental stiffness value for the presence of PE was 4.95 kPa. The variates such as GA, BMI and maternal age did not reach statistical significance as predictors of the placental stiffness value. (Fig. 4; Table 3).

Fig. 2
figure 2

Box and whisker plots of the placental ADC and μdiff values in the control and PE groups. ADC; μdiff DWI-based shear modulus, PE preeclampsia

Full size image
Table 2 Comparison of relevant parameters in the normal and PE groups
Full size table
Fig. 3
figure 3

ROC curves of the μdiff and ADC values. ADC; μdiff DWI-based shear modulus, ROC receiver-operating characteristic

Full size image
Fig. 4
figure 4

Placental stiffness measurement distribution at varying gestational ages in the control and PE groups. GA at MRI scan (week); μdiff, DWI-based shear modulus, PE preeclampsia

Full size image
Table 3 Effect of PE and common factors on placental stiffness measured by vMRE (μdiff) calculated using multiple regression analysis
Full size table

Correlation between placental stiffness and perinatal outcomes

We grouped the study population based on the optimized cut-off value for placental stiffness (e.g., Group A = placental stiffness < 4.95 kPa; Group B = placental stiffness ≥ 4.95 kPa) and evaluated their respective pregnancies and neonatal outcomes. We observed that Group B had substantially lower birth weight and GA at the time of delivery (2340 ± 958 g and 34.92 ± 4.18 weeks, respectively) compared Group A (3265 ± 580 g and 38.11 ± 1.52 weeks, respectively) (Table 4).

Table 4 Neonatal outcomes using optimized cut-off value of placental stiffness
Full size table

Discussion

In this study, we determined that the mean placental stiffness was remarkably higher in PE pregnancies than in controls, and was not affected by maternal age or GA, which was in line with previous findings [19]. However, contrasting perspectives exist in the literature; for instance, Liu et al. observed that the placental stiffness value was the lowest in 26 weeks and exhibited an upward trend from 26 to 36 weeks [20]. PE is described as an excessive maternal inflammatory response to a dysfunctional placenta or the vascular load of pregnancy itself. In normal pregnancy, placentation occurs by trophoblast invasion of the maternal spiral arteries, resulting in low-resistance, high-flow maternal uteroplacental circulation [21, 22]. In PE pregnancies, trophoblast invasion of the maternal spiral arteries is impaired, resulting in reduced placental perfusion, which creates a hypoxic environment in the placenta [21, 23]. Hypoxia stimulates collagen deposition, vascular fibrin deposition, and fibrosis, leading to higher tissue stiffness [24]. PE placentas exhibit injuries such as placental vascular lesions and vesicular fibrosis, syncytial knots and microcalcifications, and perivillous fibrin deposition and villous infarcts. These findings explain the increased placental stiffness seen in PE [25,26,27].

ADC has been reported to be lower in abnormal placentas than in normal placentas. Restricted diffusion may result from reduced gas exchange area in PE-complicated placentas [28, 29]. The present study suggested that the ADC value was remarkably lower in PE placentas than in normal placentas. The ROC curves of ADC and μdiff values were compared. The results revealed that μdiff imaging performed better than ADC in PE placentas. Accordingly, compared with ADC, μdiff is seemingly more accurate than DWI sequences in assessing dysfunctional placentas in clinical practice. Despite the widespread use of ADC, its real status of diffusivity may be limited by the monoexponential model because other factors, such as microcirculation and diffusion of water molecules, can easily interfere with it [30]. Intravoxel incoherent motion(IVIM) sequences have been developed to overcome these disadvantages using multiple b values, but the acquisition time is considerably prolonged [31, 32].

In addition, we observed that neither GA nor maternal age had noteworthy correlation with the μdiff value at MRI, consistent with the previous researches [9, 17, 33]. However, Spiliopoulos M et al. found that BMI was the only statistically significant predictor of the model with stiffness decreasing with higher BMI (inverse correlation) [9]. In their research, the BMI of pregnancies was 32.5 ± 1.4 and 34.7 ± 1.6 in the control and PE group. While in our study, the BMI of pregnancies was 21.8 ± 3.3 and 23.0 ± 3.5 in the control and PE group. BMI (calculated as weight in kilograms divided by height in meters squared) was classified in the WHO definition (underweight, < 18.5; reference weight, 18.5–24.9; overweight, 25.0–29.9; obese, ≥ 30.0) [34]. In China, pre-pregnancy BMI was categorized using the Chinese cutoffs, which are slightly different than the WHO definition (underweight, < 18.5; reference weight, 18.5–23.9; overweight, 24–27.9; obese, ≥ 28) [35].The contradiction between our results could be attributed to the individual differences and the study population selection.

The limitation to this work is that the relationship between μdiff and GA needs to be further investigated with a larger sample size. Studies on placental elastography are limited. Therefore, whether increased stiffness is seen in only PE or may be present with other causes of placental insufficiency, such as chronic hypertension, is unknown.

Conclusions

Compared with healthy pregnancies, placentas of PE pregnancies are stiffer. Placental vMRE was demonstrated to be more reliable than ADC in differentiating between normal and PE placentas. Placental stiffness is not affected by GA. It is also more likely to be associated with poor perinatal outcomes such as lower birth weight and earlier GA at delivery.