Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Nov 1.
Published in final edited form as: Am J Obstet Gynecol. 2024 Feb 27;231(5):552.e1–552.e13. doi: 10.1016/j.ajog.2024.02.291

Placental differences between severe fetal growth restriction and hypertensive disorders of pregnancy requiring early preterm delivery: Morphometric analysis of the villous tree supported by artificial intelligence

Anna JACOBS 1,*, Saif I AL-JUBOORI 2,*, Evgenia DOBRINSKIKH 3, Matthew A BOLT 4, Mary D SAMMEL 5, Virginia LIJEWSKI 6, Miriam D POST 7, James M SMALL 8, Emily J SU 9
PMCID: PMC11347726  NIHMSID: NIHMS1971943  PMID: 38423447

Abstract

Background:

The “great obstetrical syndromes” of fetal growth restriction and hypertensive disorders of pregnancy can occur individually or be interrelated. Placental pathologic findings often overlap between these conditions, regardless of whether one or both diagnoses are present. Quantification of placental villous structures in each of these settings may identify distinct differences in developmental pathways.

Objective:

To determine how quantity and surface area of placental villi and vessels differ between severe, early-onset fetal growth restriction with absent/reversed umbilical artery Doppler indices, hypertensive disorders of pregnancy, and the two conditions combined among subjects with disease severity warranting early preterm delivery. We hypothesize that trajectories of placental morphogenesis diverge after a common initiating insult of deep defective placentation. Specifically, we postulate that only villi are affected in pregnancy-related hypertension, whereas both villous and vascular structures are proportionally diminished in severe fetal growth restriction, with no additional effect when hypertension is concomitantly present.

Study design:

In this retrospective cohort study, paraffin-embedded placental tissue was obtained from four groups: [1] Severe fetal growth restriction with absent/reversed umbilical artery end-diastolic velocities and hypertensive disorders of pregnancy, [2] Severe fetal growth restriction with absent/reversed umbilical artery Doppler indices and no hypertension,[3] Gestational age-matched, appropriately grown pregnancies with hypertensive disease, and [4] Gestational age-matched, appropriately grown pregnancies without hypertension. Dual immunohistochemistry for cytokeratin-7 (trophoblast) and CD34 (endothelial cells) was performed followed by artificial intelligence-driven morphometric analyses. Number of villi, total villous area, number of fetoplacental vessels, and total vascular area across villi within a uniform region of interest were quantified. Quantitative analyses of placental structures were modeled with linear regression.

Results:

Placentas from pregnancies complicated by hypertensive disorders of pregnancy exhibited significantly fewer stem villi (−282 stem villi, 95% CI: [−467, −98], p<0.01), smaller stem villous area (−4.3 mm2, 95% CI: [−7.3, −1.2], p<0.01), and fewer stem villous vessels (−4967 stem villous vessels, 95% CI [−8501, −1433], p<0.01), with no difference in total vascular area. In contrast, placental abnormalities in severe growth restriction were limited to terminal villi, with global decreases in number of villi (−873 terminal villi, 95% CI: [−1501, −246], p<0.01), villous area (−1.5 mm2, 95% CI: [−2.7, −0.4], p<0.01), number of blood vessels (−5165 terminal villous vessels, 95% CI: [−8201, −2128], p<0.01), and vascular area (−0.6 mm2, 95% CI: [−1.1, −0.1], p=0.02). The combination of hypertension and growth restriction had no additional effect beyond the individual impact of each state.

Conclusion:

Pregnancies complicated by hypertensive disorders of pregnancy exhibited defects in the stem villi only, whereas placental abnormalities in severely growth restricted pregnancies with absent/reversed umbilical artery end-diastolic velocities were limited to the terminal villi. There were no significant statistical interactions with the combination of growth restriction and hypertension, suggesting that distinct pathophysiologic pathways downstream of the initial insult of defective placentation are involved in each entity and do not synergize to result in more severe pathologic consequences. Delineating mechanisms underlying the divergence in placental development after a common inciting event of defective deep placentation may shed light on new targets for prevention or treatment.

Keywords: Artificial intelligence, Cytokeratin-7, CD34, Doppler velocimetry, Maternal vascular malperfusion, Morphometry, Placental vasculature, Placental villi, Preeclampsia, Umbilical artery

Tweetable statement:

Differences in placental villous and vascular morphometry between severe, early-onset fetal growth restriction and hypertensive disorders of pregnancy were revealed by artificial intelligence.

Introduction

The term “great obstetrical syndromes” incorporates the concept that many pregnancy-related conditions are not discrete entities, but rather syndromes that arise from a common trigger – specifically, defective deep placentation.1–4 Depending on the response of the maternal-fetal unit, various clinical phenotypes may ensue, including fetal growth restriction (FGR), preeclampsia (PE), PE with FGR, placental abruption, second trimester loss, preterm labor with intact membranes, or preterm premature rupture of membranes.1,2 The implications of these conditions, especially those requiring early preterm delivery (<34 weeks’ gestation), include increased risks for perinatal death, medical complications, adverse neurodevelopmental sequelae, and long-term health consequences related to developmental origins of adult disease.5,6

Maternal hypertensive disorders of pregnancy (HDP) are known risk factors for development of severe FGR.7 On the other hand, FGR can also herald impending PE.8,9 Several studies have investigated placental pathologic findings in FGR with and without HDP, or FGR with and without abnormal umbilical artery Doppler velocimetry. In general, lesions consistent with maternal vascular malperfusion are more common in the presence of HDP and/or umbilical artery Doppler abnormalities.10–14 Furthermore, incorporation of the presence or absence of maternal vascular malperfusion with clinical signs and biomarkers allows for improved categorization of obstetrical syndrome subsets.15

Although consensus definitions of placental lesions exist,16,17 histopathologic diagnoses rely heavily on qualitative assessment. Quantitative analyses, may further strengthen correlation between placental structure and function, providing additional insight into pathophysiologic processes that underlie the individual “great obstetrical syndrome” phenotypes.

Placental morphometry was first described nearly 60 years ago.18,19 These and subsequent studies investigated how morphometric parameters vary with HDP, FGR, or a combination of the two.18–25 To our knowledge, quantitative analyses in early-onset PE, early-onset FGR, and both conditions combined have only been reported once, demonstrating abnormal placental morphometry in all groups.26 However, the vast majority of these FGR subjects, regardless of the presence or absence of PE, either had normal umbilical artery Doppler indices or did not undergo Doppler interrogation,26 suggesting heterogeneity in the severity of FGR subjects.

Our objective was to determine how placental villous and vascular properties differ among the closely related “great obstetrical syndrome” phenotypes that require early preterm delivery – namely, severe fetal growth restriction with absent or reversed umbilical artery end-diastolic velocities (FGRAREDV), severe HDP, or the two conditions combined. We hypothesize that quantification of key placental structures will identify divergent trajectories of placental morphogenesis after a common initiating insult of defective deep placentation. Four groups were compared: [1] FGRAREDV with HDP (FGRAREDV/HDP); [2] FGRAREDV without HDP (FGRAREDV); [3] Gestational age (GA)-matched, appropriately grown (AGA) pregnancies with hypertension (HDP); and [4] GA-matched, AGA pregnancies without FGRAREDV or HDP (Control). Number of villi, total villous area, number of villous vessels, and total vascular area were quantified using unbiased, artificial-intelligence (AI)-driven morphometric image analyses. We then examined how these properties varied with FGRAREDV, HDP, and FGRAREDV/HDP.

Materials and Methods

Study Design:

Pathology specimens were identified through the Cerner CoPath Pathology Information System for this retrospective cohort study. Potential FGRAREDV cases were identified first, followed by chart review to narrow the pool to those with persistent umbilical artery AREDV. This demonstrated that a reverse chronology of 8 years was required to achieve our sample size goal in the FGRAREDV and FGRAREDV/HDP cohorts. Equal numbers of GA-matched HDP and Control cases were identified. While we acknowledge that Control subjects experienced spontaneous preterm delivery due to pathology that may be part of the “great obstetrical syndromes” spectrum, we opted to not include a cohort of uncomplicated, full-term pregnancies as we reasoned that GA carried a substantially stronger effect on placental development than spontaneous preterm delivery. Presence or absence of FGRAREDV, HDP, or both, were considered exposure, and placental morphometric parameters were the outcomes.

Subject Selection:

This study was approved by the Colorado Multiple Institutional Review Board (COMIRB; #21-2496) and included individuals with singleton pregnancies who delivered at the UCHealth Anschutz Medical Campus between 2013-2021 and had archived placental tissue. All FGRAREDV subjects met the Delphi Consensus Criteria27,28 for early FGR, exhibited persistent umbilical artery AREDV, and underwent medically indicated preterm births at significantly preterm GAs. Subjects in the remaining cohorts were GA-matched to FGRAREDV subjects within 10 days. Concomitant hypertension in the setting of FGR (FGRAREDV/HDP) was defined as the presence of PE with severe features (n=6), superimposed PE with severe features (n=2), or chronic hypertension (n=3) as defined by the American College of Obstetricians and Gynecologists (ACOG).29,30 Specifically, PE with severe features or superimposed PE with severe features was defined as systolic blood pressure of 160 mm Hg or more, or diastolic blood pressure of 110 mm Hg or more on two occasions at least four hours apart, platelet count less than 100 x 109/L, impaired liver function with twice the upper limit of liver enzyme concentrations, renal insufficiency with a serum creatinine concentration of more than 1.1 mg/dL or a doubling, pulmonary edema, or new-onset headache unresponsive to medication.30 Chronic hypertension was defined as systolic blood pressure of 140 mm Hg or more, diastolic blood pressure of 90 mm Hg or more, or both, on two occasions at least four hours apart, which was diagnosed prior to pregnancy or before 20 weeks of gestation.29 Control subjects had birthweights between the 10th and 90th percentile for GA and experienced spontaneous preterm delivery, whereas HDP subjects underwent indicated preterm delivery secondary to PE with severe features (n=3) or superimposed PE with severe features (n=9). Exclusion criteria for all cohorts included conception via in vitro fertilization, diabetes, antiphospholipid antibody syndrome, history of thromboembolism, substance use during pregnancy, chorioamnionitis, fetal infection, anomaly, or aneuploidy/genetic syndrome.

Placental pathology:

Per institutional protocol, 2 to 4 full-thickness areas of placental parenchyma that included non-lesional and lesional areas were selected for histologic analysis. Samples were formalin-fixed, paraffin-embedded, and sectioned at 5 μm intervals. All histologic examinations were performed by a placental pathologist (MP), and lesions were categorized using the Amsterdam Placental Workshop Group Consensus Statement.16 The block chosen for analyses was non-lesional, whenever possible, in order to best represent areas of functional placenta.

Placental histologic automated image analyses:

Hematoxylin and eosin (H&E) stained slides of placental tissue (one per case) were chosen by a placental pathologist (MP) who was blinded to the clinical diagnoses. Sections were prepared for dual immunostaining of CD34 (Abcam, Waltham, MA; Rabbit mAb81289; Clone EP373Y; 1:2000 dilution) and cytokeratin-7 (Abcam; Rabbit mAb181598; Clone EPR17078; 1:2000 dilution). Antigen retrieval was performed with Borg Decloaker Solution pH 9.5 (Biocare Medical, Concord, CA) for 10 minutes at 110° C (NxGen Decloaking Chamber™, Biocare Medical) with a 10-minute ambient cool-down. Sequential dual immunostaining was performed on the Ventana Benchmark XT automated slide-staining system (Roche, Indianapolis, IN), with primary incubations for 32 minutes at 37° C for each antibody. CD34 was visualized with UltraView Universal DAB (Roche), while cytokeratin-7 was visualized with Ultraview AP Red (Roche). All sections were counterstained in Harris hematoxylin for 2 minutes, blued in 1% ammonium hydroxide, dehydrated in graded alcohols, cleared in xylene, and coverglass mounted using synthetic resin. Negative controls omitting primary antibody confirmed immunostaining specificity.

Images were acquired with the Aperio CS2 Digital Pathology Slide Scanner (Leica) at a 40x magnification (ED). Regions of interest were chosen by a placental pathologist (MP) within the central 2/3 of the parenchyma to include at least one stem villus surrounded by terminal villi measuring an average of 6 x 107 mm2 in area using ImageScope software (Leica Biosystems). After evaluation of tissue integrity and staining, samples were analyzed using the AI-driven Visiopharm™ platform (Hoersholm, Denmark). Design, development, and tissue analyses using Analysis Protocol Packages (APPs) were performed in a blinded fashion (SIA). Manual inspection for quality control of automated image analyses was performed (SIA and ED). The scanned sections in “.svs” format and the corresponding regions of interest were imported into Visiopharm™. Regions of interest were analyzed through a series of 11 custom-made APPs using the Decision Forest (80% accuracy) machine learning method that utilizes RGB bands as features and thresholding methods. At least 12 random tissue sections were used to train the algorithm. The designed APPs were used to calculate total annotation area, number of villi, individual villus areas, location references for each villus, number of blood vessels per villus, total blood vessel area per villus, and location references for blood vessels relative to their corresponding villi (Figure 1).

Figure 1. Identification and quantification of placental villi and vasculature.

Figure 1.

Open in a new tab

(A) The slide was imaged with a 40x objective and scanned. The annotated region of interest (yellow box) was selected by the placental pathologist, and the total annotation area was calculated. (B) Placental villi were recognized and outlined (yellow labels) using a series of designed protocols within the Visiopharm™ platform. The total number of villi (3647 total villi within this representative region of interest) and areas of each individual villus were calculated. (C) Magnification of the cyan box within the overall region of interest shows individual villi outlined by yellow lines. Within each detected villus, fetoplacental blood vessels were identified and delineated by the magenta line using a separate series of designed protocols in Visiopharm™. Total blood vessel area in each villus was calculated. (D) Blood vessels were separated using yet another series of designed protocols in Visiopharm™, and number of blood vessels in each villus was quantified.

Initial data included 197,308 measures of number of villi, villous area, blood vessel count, and blood vessel area. Structures identified as villi measuring less than 300 μm2 were more consistent with cell debris and thereby excluded (n=1527, 0.77% of original sample). Four villi that had zero blood vessels but non-zero blood vessel areas were also removed, accounting for 0.002% of the total sample. As villi measuring less than 75 μm in diameter are considered mature intermediate and terminal villi,31 we utilized a cutoff of 4415 μm2 (πr2: π*37.52) in area to separate stem from terminal villi. After cleaning, an average of 2887 terminal villi and 952 stem villi were analyzed per subject.

Statistical analysis:

Sample size calculations were conducted with G*Power 3.1.9.7.32 Assuming a two-factor ANOVA with an interaction study design, 80% power, a type I error rate of 5%, and an average terminal villous volume of 150 cm3 with standard deviation (SD) of 58.8 cm3 for control subjects based on a manuscript that reported placental morphometric characteristics among similar participants,26 we calculated that a sample size of 11 per group would be necessary to detect an interaction effect size of 0.43 SD (“Cohen’s f”37) or larger. This effect size corresponds to a decrease in minimum detectable villous volume of 71.1 cm3 for FGRAREDV/HDP and a minimum difference in terminal villous area of 1.18 m2 when compared to controls. Both these values exceeded those as previously reported,26 demonstrating the feasibility of the proposed sample size. In anticipation of potential staining issues, this study attempted to obtain 15 subjects per group.

For quantitative analyses of stem and terminal villous components, all measures were modeled with linear regression models after distribution of outcomes were found to be adequately Gaussian, as determined by visualization of density plots and histograms. R statistical software (version 4.2.2) was used for all statistical analyses in conjunction with the following R packages: magittr, pwr, arsenal, here, broom, tidyverse, readxl, and kableExtra.33–42

Results

Study population characteristics:

After excluding subjects with suboptimal immunohistochemical staining quality, final analyses included 51 participants: 14 Control, 12 HDP, 14 FGRAREDV, and 11 FGRAREDV /HDP subjects (Table 1). As anticipated, individuals with FGRAREDV had lower observed neonatal birthweights and birthweight percentiles for GA and displayed persistent umbilical artery AREDV. When further stratifying the AGA cohort by the presence or absence of hypertension, HDP pregnancies with AGA fetuses appeared to have lower birthweights and birthweight percentiles for GA than the Control cohort. However, the birthweight percentiles of the HDP group confirm that these neonates did not meet ACOG criteria for FGR.

Table 1.

Study subject population characteristics.

Characteristic Control (N=14) HDP (N=12) FGRAREDV (N=14) FGRAREDV/HDP (N=11)
Maternal age (years)
 Mean (SD) 27 (8) 30 (5) 27 (5) 32 (6)
 Median (IQR) 26 (19, 34) 29 (26, 31) 26 (23, 30) 34 (31, 34)
Parity
 Mean (SD) 2 (2) 1 (1) 1 (1) 1 (1)
 Median (IQR) 1 (0, 2) 1 (1, 2) 0 (0, 1) 1 (0, 2)
GA at delivery (weeks)
 Mean (SD) 30.20 (3.19) 28.97 (2.87) 30.51 (3.32) 27.60 (3.12)
 Median (IQR) 30.21 (27.28, 33.31) 29.42 (26.49, 31.32) 30.56 (27.14, 33.84) 26.70 (24.92, 30.14)
Maternal race/ethnicity
 White 6 (43%) 4 (33%) 7 (50%) 4 (36%)
 Hispanic or Latina 3 (21%) 6 (50%) 5 (36%) 2 (18%)
 Black or African-American 2 (14%) 1 (8.3%) 0 (0%) 2 (18%)
 Other 3 (21%) 1 (8.3%) 2 (14%) 3 (27%)
Earliest pregnancy BMI (kg/m2)
 Mean (SD) 26.0 (4.8) 35.2 (9.5) 28.2 (6.7) 29.1 (4.6)
 Median (IQR) 24.9 (24.0, 25.8) 35.2 (31.9, 38.6) 26.4 (23.5, 31.3) 28.8 (25.9, 32.7)
 Unknown 6 10 2 3
Umbilical artery Doppler
 Absent end-diastolic velocities 0 (0%) 0 (0%) 10 (71%) 5 (45%)
 Reversed end-diastolic velocities 0 (0%) 0 (0%) 4 (29%) 6 (55%)
 Not performed 14 (100%) 12 (100%) 0 (0%) 0 (0%)
Oligohydramnios 0 (0%) 1 (8.3%) 2 (14%) 0 (0%)
Neonatal sex
 Female 2 (14%) 7 (58%) 7 (50%) 5 (45%)
 Male 12 (86%) 5 (42%) 7 (50%) 6 (55%)
Neonatal birth weight (g)
 Mean (SD) 1,644 (609) 1,150 (354) 873 (413) 720 (373)
 Median (IQR) 1,585 (1,201, 2,082) 1,230 (849, 1,442) 775 (564, 1,255) 515 (428, 980)
Neonatal birthweight %ile for GA
 Mean (SD) 66 (21) 34 (24) 4 (5) 7 (8)
 Median (IQR) 62 (50, 88) 25 (16, 45) 2 (0, 4) 3 (1, 10)
Open in a new tab

Data are presented as mean (standard deviation), median (interquartile range), or count (percentage).

HDP, hypertensive disease of pregnancy; FGRAREDV, fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry; FGRAREDV/HDP, combination of hypertensive disease of pregnancy and fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry; GA, gestational age

Placental pathology:

Table 2 summarizes placental histopathologic findings for all four cohorts, and Figure 2 depicts both representative H&E and dual-immunostained images for cytokeratin-7 and CD34 from each cohort. Placental weight and respective percentile for GA appeared to be lower in those with HDP, FGRAREDV, or both combined. Lesions of maternal vascular malperfusion were uncommon in Controls but highly prevalent in the HDP, FGRAREDV, and FGRAREDV/HDP cohorts. Of maternal vascular malperfusion lesions, accelerated villous maturation was most common. Further demonstrating the severity of pathology in our FGRAREDV cohort, we also observed higher rates of fetal vascular malperfusion in these placentas.

Table 2.

Placental pathologic findings.

Characteristic Control (N=14) HDP (N=12) FGRAREDV (N=14) FGRAREDV/HDP (N=11)
Placental weight (grams)
 Mean (SD) 310 (91) 231 (56) 176 (84) 141 (46)
 Median (IQR) 317 (288, 342) 236 (192, 274) 148 (118, 205) 135 (111, 150)
Placental weight %ile for GA
 Mean (SD) 43 (29) 19 (21) 15 (23) 10 (3)
 Median (IQR) 34 (25, 50) 10 (10, 19) 10 (4, 10) 10 (10, 10)
Any maternal vascular malperfusion 2 (14%) 12 (100%) 11 (79%) 10 (91%)
 Presence of infarcts 0 (0%) 5 (42%) 2 (14%) 7 (64%)
 Retroplacental hemorrhage 0 (0%) 5 (42%) 5 (36%) 3 (27%)
 Accelerated villous maturation 2 (14%) 7 (58%) 8 (57%) 7 (64%)
 Decidual arteriopathy 0 (0%) 2 (17%) 1 (7.1%) 6 (55%)
Any fetal vascular malperfusion 0 (0%) 3 (25%) 11 (79%) 2 (18%)
 Thrombosis 0 (0%) 2 (17%) 0 (0%) 0 (0%)
 Avascular villi 0 (0%) 0 (0%) 5 (36%) 0 (0%)
 Intramural fibrin deposition 0 (0%) 1 (8.3%) 7 (50%) 1 (9.1%)
 Villous stromal vascular karyorrhexis 0 (0%) 0 (0%) 2 (14%) 1 (9.1%)
 Stem vessel obliteration 0 (0%) 0 (0%) 2 (14%) 1 (9.1%)
Maternal inflammatory response 3 (21%) 4 (33%) 10 (71%) 3 (27%)
Fetal inflammatory response 4 (29%) 0 (0%) 4 (29%) 0 (0%)
Villitis of unknown etiology 1 (7.1%) 0 (0%) 3 (21%) 0 (0%)
Open in a new tab

Data are presented as mean (standard deviation), median (interquartile range), or count (percentage).

HDP, hypertensive disease of pregnancy; FGRAREDV, fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry; FGRAREDV/HDP, combination of hypertensive disease of pregnancy and fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry; GA, gestational age

Figure 2. Representative placental histologic images.

Figure 2.

Open in a new tab

Hematoxylin and eosin (H&E) photomicrographs with corresponding dual immunohistochemical-stained images (CK7 in red, CD34 in brown) are shown for each of the four study cohorts. All images were taken at 04x magnification.

We then quantified number of villi, villous area, number of vessels, and vascular area between the four study cohorts. As stem and terminal villi serve different functions within the placenta, we further segregated our morphometric analyses into these two categories (Table 3).

Table 3.

Stem and terminal villi morphometric measures

Control N=14 HDP N=12 FGRAREDV N=14 FGRAREDV/HDP N=11
Stem Villi
Number of villi
 Mean (SD) 1174 (332) 808 (384) 982 (321) 788 (252)
Total villous area (mm2)
 Mean (SD) 18 (5) 13 (6) 16 (6) 13 (4)
Number of blood vessels
 Mean (SD) 19,017 (5684) 12,855 (6168) 16,049 (8395) 12,336 (3130)
Total blood vessel area (mm2)
 Mean (SD) 4.6 (1.7) 3.8 (2.2) 4.1 (1.5) 3.5 (1.2)
Terminal Villi
Number of villi
 Mean (SD) 3280 (832) 3363 (1065) 2211 (1053) 2728 (1520)
Total villous area (mm2)
 Mean (SD) 6 (2) 6 (2) 5 (2) 5 (3)
Number of blood vessels
 Mean (SD) 15,366 (5466) 16,272 (5266) 9603 (4942) 11,837 (6148)
Total blood vessel area (mm2)
 Mean (SD) 2.3 (0.8) 2.2 (0.8) 1.6 (0.8) 1.8 (1.1)
Open in a new tab

HDP, hypertensive disease of pregnancy; FGRAREDV, fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry; FGRAREDV/HDP, combination of hypertensive disease of pregnancy and fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry

Quantitative stem villus and vessel analyses:

Linear regression of data from image analyses demonstrated that cumulatively, number of stem villi did not significantly differ by AGA or FGRAREDV status (Table 4). Placentas from HDP pregnancies, however, displayed significantly fewer stem villi (−282, 95% CI: [−467, −98], p<0.01) than those without hypertension (Table 4). The combination of FGRAREDV/HDP did not further alter number of stem villi beyond the effect of HDP alone (Table 5). There was also a significant reduction in stem villous area for those with HDP (−4.3 mm2, 95% CI: [−7.3, −1.2], p<0.01), whereas stem villous area did not otherwise differ by FGRAREDV status nor by the combination of FGRAREDV/HDP (Tables 4 and 5).

Table 4.

Multivariable linear regression for stem villi and terminal villi parameters

Stem Villi Terminal Villi
Outcome Variable Estimate CI p Estimate CI p
Number of villi Effect of FGRAREDV −114 (−298, 70) 0.22 −873 (−1501, −246) < 0.01
Effect of HDP −282 (−467, −97) < 0.01 294 (−336, 925) 0.35
Total villous area (mm2) Effect of FGRAREDV −1.0 (−4.1, 2.0) 0.51 −1.5 (−2.7, −0.4) < 0.01
Effect of HDP −4.3 (−7.3, −1.2) < 0.01 0.2 (−1.0, 1.3) 0.77
Number of blood vessels Effect of FGRAREDV −1865 (−5383, 1653) 0.29 −5165 (−8201, −2128) < 0.01
Effect of HDP −4967 (−8501, −1433) < 0.01 1554 (−1497, 4605) 0.31
Total blood vessel area (mm2) Effect of FGRAREDV −0.4 (−1.3, 0.5) 0.41 −0.6 (−1.1, −0.1) 0.02
Effect of HDP −0.7 (−1.7, 0.2) 0.13 0.1 (−0.4, 0.6) 0.79
Open in a new tab

Separate multivariable linear regression models were performed for stem villi and terminal villi. The “Effect of FGRAREDV” estimate (β1) represents the mean difference between subjects with FGRAREDV as compared to pregnancies with fetuses that were appropriately grown for gestational age (AGA). The “Effect of HDP” estimate (β2) represents the mean difference between subjects with HDP and those without hypertension. Model intercept (β0) not shown.

HDP, hypertensive disease of pregnancy; FGRAREDV, fetal growth restriction with absent or reversed umbilical artery Doppler velocimetry

Table 5.

Effect of the combination (interaction) of FGRAREDV/HDP on stem and terminal villi outcomes

Stem Villi Terminal Villi
Outcome Estimate CI p Estimate CI p
Number of villi 171 (−199, 541) 0.36 434 (−836, 1703) 0.50
Total villous area (mm2) 2.6 (−3.5, 8.8) 0.39 0.9 (−1.4, 3.1) 0.45
Number of blood vessels 2449 (−4664, 9561) 0.49 1328 (−4830, 7486) 0.67
Total blood vessel area (mm2) 0.3 (−1.6, 2.2) 0.77 0.3 (−0.7, 1.3) 0.50
Open in a new tab

The estimates above are linear regression (β3) coefficients for the statistical interactions between FGRAREDV and HDP status on the outcomes listed in the rows. Separate multivariable linear regression models were performed for stem villi and terminal villi. The intercept and main effects for these models are not shown (β0, β1, and β2). Since no interactions were statistically significant, final models (fit only with main effects) are reported in Table 4.

Number of blood vessels within the stem villi were also similar between AGA and FGRAREDV placentas, although again, placentas with HDP exhibited significantly fewer stem villous vessels as compared to those without HDP (−4967, 95% CI: [−8501, −1433], p<0.01) (Table 4). Despite fewer vessels, there were no significant differences in total stem villous blood vessel area regardless of FGRAREDV, HDP, or their combination (Tables 4 and 5). These data suggest that in the setting of HDP, a compensatory response such as vasodilatation may occur, although artifactual dilation in response to tissue processing cannot be entirely excluded. In contrast, FGRAREDV did not demonstrate significant differences in stem villous metrics, regardless of whether hypertension was present (Tables 4 and 5), indicating that placental abnormalities underlying development of FGRAREDV are likely not at the level of stem villi and vessels.

Quantitative terminal villus and vessel analyses:

Placentas with FGRAREDV demonstrated significantly fewer terminal villi than AGA placentas (−873, 95% CI: [−1501, −246], p<0.01) (Table 4). This further translated into significant reductions in total villous area (−1.5 mm2, 95% CI: [−2.7, −0.4], p<0.01), number of blood vessels (−5165, 95% CI: [−8201, −2128], p<0.01), and total blood vessel area (−0.6 mm2, 95% CI: [−1.1, −0.1], p=0.02). Conversely, the presence of HDP did not alter these parameters (Table 4), nor did the combination of FGRAREDV/HDP (Table 5), suggesting that an inadequate surface area for maternal-fetal exchange is associated with FGRAREDV.

Comment

Principal Findings:

We found that pregnancies complicated by HDP exhibited defects in the stem villi only, with fewer stem villi, reduced stem villous area, and fewer stem villous vessels, although blood vessel area was similar to that seen in placentas without HDP. In contrast, placental abnormalities in FGRAREDV pregnancies were limited to the terminal villi, with decreases in number of terminal villi, terminal villous area, number of terminal villous vessels, and terminal villous blood vessel area. There were no statistically significant interactions with the combination of HDP and FGRAREDV as compared to HDP or FGRAREDV alone, suggesting that the distinct pathophysiologic pathways leading to HDP or FGRAREDV do not synergize to result in more severe pathologic consequences.

Results in the context of what is known:

It is well-established that placental lesions consistent with both maternal and fetal vascular malperfusion are more common in pregnancies complicated by FGRAREDV, PE with severe features, or in those with both conditions.10–14 Defective deep placentation is likely a common inciting factor, as suggested by the similar appearance of placental beds in individuals with PE, FGR, and PE with FGR, containing numerous myometrial junctional zone spiral arteries that have not undergone proper transformation.1,2,43–47

Placental morphometric studies spanning several decades demonstrate that hypertension, FGR, or concomitant hypertension and FGR each carry structural consequences. Most show some degree of reduction in villous and vascular volumes and surface area with HDP, FGR, or both.18–24 However, these publications included heterogeneous study populations both within and between studies, ranging in severity of disease and from preterm to late-term deliveries. Investigations into morphometry specifically related to abnormal umbilical artery Doppler indices have also demonstrated reductions in terminal villous and vascular structures within these villi, although again, clinical phenotypes varied.25,48,49 To our knowledge, only one study by Egbor and colleagues has specifically interrogated placental morphometry in early-onset PE, early-onset FGR, and the combination of the two.26 Those authors found that as compared to a GA-matched control population without PE or FGR, there were significant decreases in terminal villous surface area in PE, FGR, and their combination. However, despite necessitating preterm delivery, 5/7 of the FGR cohort and 5/9 of the PE/FGR cohort either did not have umbilical artery Doppler velocimetry performed or exhibited normal indices. Additionally, standard histopathologic findings were not reported. Together, this suggests heterogeneity in the phenotypes of interest. Furthermore, in the morphometry studies cited, measurements were calculated using random sampling and counting of chance events between planes, lines, or points, as compared to actual measurements of all individual structures within a region of interest.

Our overarching goal is to understand the underpinnings of how well-defined, clinical phenotypes -- specifically, HDP, early-onset FGRAREDV, or the two conditions combined -- develop after the common insult of defective deep placentation. In contrast to standard morphometry principles, we utilized an AI-based image analysis platform trained to quantify villous and vascular structures within a uniform, non-lesional region of interest.

Development of stem villi is temporally related to formation of immature intermediate villi, which begins as early as 8 weeks’ gestation, and expansion of villous stems requires immature intermediate villi.31 Our data showing more proximal villus abnormalities in hypertensive pregnancies suggest that the impact of defective placentation occurs earlier in pregnancies that ultimately manifest HDP. Furthermore, there may be some compensatory vasodilation, as placentas complicated by HDP show similar blood vessel area despite fewer vessels.

In FGRAREDV, the lack of change in stem villi but global decrement in terminal villous structures suggest an impairment of villous branching, which normally occurs throughout the second half of gestation.31 This defect likely results in decreased surface area for maternal-fetal exchange. Whether this is the cause of FGRAREDV or an adaptation to decreased nutrient/oxygen availability from defective deep placentation remains unknown. However, our findings suggest that perturbations in normal placental developmental processes are occurring later in gestation when terminal villi expand in FGRAREDV. Further supporting the distinct timeframes of the potential insults in HDP and FGRAREDV is the fact that the combination of the two does not further impact villous and vascular structures.

Clinical and research implications:

Despite substantial scientific advances, the field of obstetrics remains limited in its ability to prevent or treat pregnancy-related conditions. One obstacle is the assumption that a specific disorder occurs due to the same pathological mechanisms across all individuals. Instead, the field of obstetrics is evolving toward the concept of “great obstetrical syndromes,” where a single root cause of defective deep placentation, modified by individual factors, may give rise to diverse phenotypes. This conceptual shift may facilitate characterization of the various factors that cause seemingly disparate syndromes.

Standardization in placental pathology has substantially aided our understanding of potential insults resulting in a suboptimally functioning placenta. Yet, it also highlights the pathophysiologic heterogeneity of specific pregnancy diagnoses. Beyond commonalities between pregnancies with HDP and FGRAREDV, studies comparing placental pathologic findings between spontaneous and indicated preterm birth also demonstrate overlap in findings, with different proportions of both cohorts displaying maternal vascular malperfusion and inflammatory lesions.50,51

As an adjunct to standard histologic evaluation, AI-driven image analysis has been shown to be a highly accurate, diagnostic tool in multiple organs and diseases.52–54 In the obstetric setting, this platform provides the opportunity for unbiased quantification of placental structures, and in contrast to standard morphometry, has the capacity to rapidly evaluate thousands of villi per image. These data, especially when applied to rigorously defined phenotypes and sub-phenotypes of the “great obstetrical syndromes,” may enhance our understanding of the various pathways that can be involved in a single syndrome. For example, in a “small for gestational age” (SGA) neonate with antenatal concern for FGR and birthweight less than the 10th percentile for GA, AI-driven morphometric analyses may show substantial decreases in terminal villous structures, suggesting a pathologically small infant. In contrast, normal quantity and area of terminal villous structures could suggest a neonate that has met his or her inherent growth potential. Expeditiously ascertaining whether the birthweight of this neonate is pathologic versus constitutional has the potential to affect pediatric care. It may also lead to models that predict risks for subsequent pregnancies.

A central question in many mechanistic-based studies is whether findings are the cause of the pathologic phenotype or compensatory responses to an insult. This fundamental missing link is one obstacle impeding development of effective interventions. While placental morphometric analysis alone is unlikely to answer this question, the extended knowledge provided by this platform has the potential to guide future investigations.

Strengths and limitations:

Strengths of this study include the focus on individuals with severe phenotypes requiring very preterm delivery and the development of novel AI-driven technology to provide quantitative placental analyses in a quick and unbiased fashion. One limitation is that only one piece of tissue was analyzed per subject. However, standard pathologic analysis of placentas includes only 2 to 4 sections of parenchyma, and for this study, the section with histology best illustrating the complete villous tree was chosen in order to ascertain the background milieu of the placenta. Another limitation is that we did not train the AI algorithm to categorize villi, and we used existing literature to identify a diameter cutoff value that segregated stem from terminal villi.31 As placental morphology inadvertently results in some structures being cut on the diagonal (versus perpendicular to each villus), this could result in a small number of structures being incorrectly categorized as stem villi.

We also acknowledge that neither the Control nor HDP cohorts have Doppler data available given the retrospective nature of this study and the standard-of-care where umbilical artery Doppler interrogation is not performed in fetuses that are not suspected to be growth restricted. Furthermore, although we attempted to rigorously define our cohorts, there may still be heterogeneity among individuals. For instance, all HDP subjects were delivered preterm due to early-onset PE with severe features or superimposed PE with severe features. However, a robust body of literature suggests that mechanistic diversity exists even in strictly defined cases of early-onset PE with severe features.55–59 Different etiologic factors predispose to early-onset PE. Even if these converge to induce cellular dysfunction (e.g. endothelial dysfunction, trophoblast stress, immune intolerance) associated with PE with severe features, the mechanisms initiating cellular damage likely differ.56 Furthermore, unbiased microarray data show that different molecular signatures exist even among individuals with severe forms of PE.55,57,58 With regard to the FGRAREDV/HDP cohort, 8 of 11 carried the diagnosis of PE with severe features or superimposed PE with severe features, and 3 had chronic hypertension requiring anti-hypertensive therapy. Six of the 11 delivered as a result of maternal indications, whereas 5 deliveries occurred secondary to fetal indications, suggesting that these pregnancies may further differ. Individuals with chronic hypertension were included in this FGRAREDV/HDP cohort as it is not uncommon for individuals with FGRAREDV at our institution to also manifest chronic hypertension without superimposed PE. A subanalysis comparing morphometric parameters between FGRAREDV/HDP with chronic hypertension and FGRAREDV/HDP with PE with severe features or superimposed PE with severe features showed no differences among any of the measures (Supplemental Table 1).

We also acknowledge that placentas from spontaneous preterm birth can exhibit maternal vascular malperfusion, as found in 2/14 control subjects.60–64 However, placental villous and vascular development are GA-dependent. Thus, we reasoned that including full-term subjects with uncomplicated pregnancies would limit our understanding of the role of HDP and FGRAREDV on placental villous structures.

Conclusions:

Using AI-driven placental morphometric analysis to quantify placental villous and vascular structures, we found that pregnancies complicated by HDP display changes primarily limited to more proximal stem villi, whereas pregnancies complicated by FGRAREDV exhibit terminal villous abnormalities, with no additional findings when HDP and FGRAREDV are concomitantly present. Future studies leveraging this technology may help identify heterogeneous pathways resulting in specific great obstetrical syndromes and guide additional research efforts to better understand mechanisms contributing to insults that ultimately result in pregnancy-related disorders.

Supplementary Material

1
2
Download video file (193.4MB, mov)
3
Download video file (45.4MB, mov)
4

AJOG at a Glance:

A. Why was this study conducted?

  • Fetal growth restriction and preeclampsia are two “great obstetrical syndromes” sharing a trigger of defective placentation yet manifesting distinct or interrelated phenotypes.

  • This study aimed to determine how placental villous and vascular properties differ between subjects with fetal growth restriction and absent/reversed umbilical artery Doppler velocimetry, severe hypertension, or both, to elucidate respective differences in placental development.

B. What are the key findings?

  • Placentas complicated by hypertension exhibited reduced stem villi, stem villous area, and stem vessels, whereas growth-restricted placentas exhibited reduced terminal villi, terminal villous vessels, and terminal villous and vascular surface areas. No additional effects were found when conditions were concomitant.

C. What does this study add to what is already known?

  • This is the first study that specifically compares placental morphometry between pregnancies with severely growth-restricted fetuses with absent or reversed umbilical artery Doppler velocimetry and those with hypertension requiring early preterm delivery.

Acknowledgements:

We are grateful to E. Erin Smith for immunohistochemical staining.

Funding source:

NIH HL119846 (EJS). The sponsor had no involvement in study design, data collection, analysis, or interpretation, writing of the manuscript, or decision to submit the article for publication.

GLOSSARY OF TERMS

Accelerated villous maturation

Presence of small or short hypermature villi for gestational age, usually accompanied by an increase in syncytial knots

Avascular villi

Terminal villi showing total loss of villous capillaries and hyaline fibrosis of the villous stroma

Decidual arteriopathy

Abnormalities of maternal spiral arteries; often associated with maternal vascular malperfusion; includes acute atherosis (fibrinoid necrosis), mural hypertrophy, and chronic perivasculitis, primarily affecting membranous arterioles, and persistent muscularization of basal plate arteries

Fetal inflammatory response

Cellular inflammatory response to amniotic fluid infection defined by neutrophils within the walls of fetal blood vessels, either in the umbilical cord or chorionic plate

Fetal vascular malperfusion

Obstructed fetal blood flow because of thrombosis and/or prolonged umbilical cord compression leading to avascular villi

Intramural fibrin deposition

Subendothelial or intramuscular fibrin deposition Maternal inflammatory response: Cellular inflammatory response to infection resulting in presence of neutrophils in the subchorial intervillous space, chorion, and/or amnion

Maternal vascular malperfusion

Defined by accelerated villous maturation: alternating areas of villous paucity and crowding with increased syncytial knots and focal intervillous fibrin; often accompanied by infarcts and decreased placental weight

Stem vessel obliteration

Marked thickening of the vessel wall with resultant obliteration of the vascular lumen

Villitis of unknown etiology

Infiltration of chorionic villous stroma by small lymphocytes and activated macrophages, often accompanied by avascular villi and pervillous fibrin

Villous stromal-vascular karyorrhexis

Three or more foci of 2 to 4 terminal villi showing karyorrhexis of fetal cells (nucleated erythrocytes, leukocytes, endothelial cells, and/or stromal cells) with preservation of surrounding trophoblast

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disclosure statement of any potential of interest: The authors report no conflict of interest.

Contributor Information

Ms. Anna JACOBS, Rocky Vista University College of Osteopathic Medicine, Parker, CO.

Dr. Saif I AL-JUBOORI, University of Colorado School of Medicine, Department of Pediatrics, Section of Neonatology, Aurora, CO.

Dr. Evgenia DOBRINSKIKH, University of Colorado School of Medicine, Department of Pediatrics, Department of Medicine, Section of Neonatology, Division of Pulmonary Sciences and Critical Care Medicine, Aurora, CO.

Mr. Matthew A BOLT, Colorado School of Public Health, Center for Innovative Design and Analysis, Department of Biostatistics and Informatics, Aurora, CO.

Dr. Mary D SAMMEL, Colorado School of Public Health, Center for Innovative Design and Analysis, Department of Biostatistics and Informatics, Aurora, CO.

Ms. Virginia LIJEWSKI, University of Colorado School of Medicine, Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, Aurora, CO.

Dr. Miriam D POST, University of Colorado School of Medicine, Department of Pathology, Aurora, CO.

Dr. James M SMALL, Rocky Vista University College of Osteopathic Medicine, Department of Biomedical Sciences, Parker, CO.

Dr. Emily J SU, University of Colorado School of Medicine, Department of Obstetrics and Gynecology, Division of Reproductive Sciences/Division of Maternal-Fetal Medicine, Aurora, CO.

References:

  • 1.Brosens I, Pijnenborg R, Vercruysse L, Romero R. The “Great Obstetrical Syndromes” are associated with disorders of deep placentation. Am J Obstet Gynecol. Mar 2011;204(3):193–201. doi: 10.1016/j.ajog.2010.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Brosens I, Puttemans P, Benagiano G. Placental bed research: I. The placental bed: from spiral arteries remodeling to the great obstetrical syndromes. Am J Obstet Gynecol. Nov 2019;221(5):437–456. doi: 10.1016/j.ajog.2019.05.044 [DOI] [PubMed] [Google Scholar]
  • 3.Di Renzo GC. The great obstetrical syndromes. J Matern Fetal Neonatal Med. Aug 2009;22(8):633–5. doi: 10.1080/14767050902866804 [DOI] [PubMed] [Google Scholar]
  • 4.Romero R. Prenatal medicine: the child is the father of the man. 1996. J Matern Fetal Neonatal Med. Aug 2009;22(8):636–9. doi: 10.1080/14767050902784171 [DOI] [PubMed] [Google Scholar]
  • 5.Barfield WD. Public Health Implications of Very Preterm Birth. Clin Perinatol. Sep 2018;45(3):565–577. doi: 10.1016/j.clp.2018.05.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Calkins K, Devaskar SU. Fetal origins of adult disease. Curr Probl Pediatr Adolesc Health Care. Jul 2011;41(6):158–76. doi: 10.1016/j.cppeds.2011.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Fetal Growth Restriction: ACOG Practice Bulletin, Number 227. Obstet Gynecol. Feb 1 2021;137(2):e16–e28. doi: 10.1097/AOG.0000000000004251 [DOI] [PubMed] [Google Scholar]
  • 8.Mitani M, Matsuda Y, Makino Y, Akizawa Y, Ohta H. Clinical features of fetal growth restriction complicated later by preeclampsia. J Obstet Gynaecol Res. Oct 2009;35(5):882–7. doi: 10.1111/j.1447-0756.2009.01120.x [DOI] [PubMed] [Google Scholar]
  • 9.Takahashi M, Makino S, Oguma K, et al. Fetal growth restriction as the initial finding of preeclampsia is a clinical predictor of maternal and neonatal prognoses: a single-center retrospective study. BMC Pregnancy Childbirth. Oct 6 2021;21(1):678. doi: 10.1186/s12884-021-04152-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ganer Herman H, Barber E, Gasnier R, et al. Placental pathology and neonatal outcome in small for gestational age pregnancies with and without abnormal umbilical artery Doppler flow. Eur J Obstet Gynecol Reprod Biol. Mar 2018;222:52–56. doi: 10.1016/j.ejogrb.2018.01.009 [DOI] [PubMed] [Google Scholar]
  • 11.Kovo M, Schreiber L, Ben-Haroush A, Wand S, Golan A, Bar J. Placental vascular lesion differences in pregnancy-induced hypertension and normotensive fetal growth restriction. Am J Obstet Gynecol. Jun 2010;202(6):561 e1–5. doi: 10.1016/j.ajog.2010.01.012 [DOI] [PubMed] [Google Scholar]
  • 12.Salafia CM, Pezzullo JC, Minior VK, Divon MY. Placental pathology of absent and reversed end-diastolic flow in growth-restricted fetuses. Obstet Gynecol. Nov 1997;90(5):830–6. doi: 10.1016/S0029-7844(97)00473-0 [DOI] [PubMed] [Google Scholar]
  • 13.Veerbeek JH, Nikkels PG, Torrance HL, et al. Placental pathology in early intrauterine growth restriction associated with maternal hypertension. Placenta. Sep 2014;35(9):696–701. doi: 10.1016/j.placenta.2014.06.375 [DOI] [PubMed] [Google Scholar]
  • 14.Freedman AA, Suresh S, Ernst LM. Patterns of placental pathology associated with preeclampsia. Placenta. Jun 12 2023;139:85–91. doi: 10.1016/j.placenta.2023.06.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Romero R, Jung E, Chaiworapongsa T, et al. Toward a new taxonomy of obstetrical disease: improved performance of maternal blood biomarkers for the great obstetrical syndromes when classified according to placental pathology. Am J Obstet Gynecol. Oct 2022;227(4):615 e1–615 e25. doi: 10.1016/j.ajog.2022.04.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Khong TY, Mooney EE, Ariel I, et al. Sampling and Definitions of Placental Lesions: Amsterdam Placental Workshop Group Consensus Statement. Arch Pathol Lab Med. Jul 2016;140(7):698–713. doi: 10.5858/arpa.2015-0225-CC [DOI] [PubMed] [Google Scholar]
  • 17.Redline RW, Roberts DJ, Parast MM, et al. Placental pathology is necessary to understand common pregnancy complications and achieve an improved taxonomy of obstetrical disease. Am J Obstet Gynecol. Feb 2023;228(2):187–202. doi: 10.1016/j.ajog.2022.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Aherne W, Dunnill MS. Morphometry of the human placenta. Br Med Bull. Jan 1966;22(1):5–8. doi: 10.1093/oxfordjournals.bmb.a070437 [DOI] [PubMed] [Google Scholar]
  • 19.Aherne W, Dunnill MS. Quantitative aspects of placental structure. J Pathol Bacteriol. Jan 1966;91(1):123–39. doi: 10.1002/path.1700910117 [DOI] [PubMed] [Google Scholar]
  • 20.Boyd PA, Scott A. Quantitative structural studies on human placentas associated with pre-eclampsia, essential hypertension and intrauterine growth retardation. Br J Obstet Gynaecol. Jul 1985;92(7):714–21. doi: 10.1111/j.1471-0528.1985.tb01454.x [DOI] [PubMed] [Google Scholar]
  • 21.Daayana S, Baker P, Crocker I. An image analysis technique for the investigation of variations in placental morphology in pregnancies complicated by preeclampsia with and without intrauterine growth restriction. J Soc Gynecol Investig. Dec 2004;11(8):545–52. doi: 10.1016/j.jsgi.2004.06.009 [DOI] [PubMed] [Google Scholar]
  • 22.Egbor M, Ansari T, Morris N, Green CJ, Sibbons PD. Pre-eclampsia and fetal growth restriction: how morphometrically different is the placenta? Placenta. Jun-Jul 2006;27(6-7):727–34. doi: 10.1016/j.placenta.2005.06.002 [DOI] [PubMed] [Google Scholar]
  • 23.Mayhew TM, Ohadike C, Baker PN, Crocker IP, Mitchell C, Ong SS. Stereological investigation of placental morphology in pregnancies complicated by pre-eclampsia with and without intrauterine growth restriction. Placenta. Feb-Mar 2003;24(2-3):219–26. doi: 10.1053/plac.2002.0900 [DOI] [PubMed] [Google Scholar]
  • 24.Mayhew TM, Wijesekara J, Baker PN, Ong SS. Morphometric evidence that villous development and fetoplacental angiogenesis are compromised by intrauterine growth restriction but not by pre-eclampsia. Placenta. Nov 2004;25(10):829–33. doi: 10.1016/j.placenta.2004.04.011 [DOI] [PubMed] [Google Scholar]
  • 25.Mitra SC, Seshan SV, Riachi LE. Placental vessel morphometry in growth retardation and increased resistance of the umbilical artery Doppler flow. J Matern Fetal Med. Sep-Oct 2000;9(5):282–6. doi: [DOI] [PubMed] [Google Scholar]
  • 26.Egbor M, Ansari T, Morris N, Green CJ, Sibbons PD. Morphometric placental villous and vascular abnormalities in early- and late-onset pre-eclampsia with and without fetal growth restriction. BJOG. May 2006;113(5):580–9. doi: 10.1111/j.1471-0528.2006.00882.x [DOI] [PubMed] [Google Scholar]
  • 27.Gordijn SJ, Beune IM, Thilaganathan B, et al. Consensus definition of fetal growth restriction: a Delphi procedure. Ultrasound Obstet Gynecol. Sep 2016;48(3):333–9. doi: 10.1002/uog.15884 [DOI] [PubMed] [Google Scholar]
  • 28.Lees CC, Romero R, Stampalija T, et al. Clinical Opinion: The diagnosis and management of suspected fetal growth restriction: an evidence-based approach. Am J Obstet Gynecol. Mar 2022;226(3):366–378. doi: 10.1016/j.ajog.2021.11.1357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.ACOG Practice Bulletin No. 203 Summary: Chronic Hypertension in Pregnancy. Obstet Gynecol. Jan 2019;133(1):215–219. doi: 10.1097/AOG.0000000000003021 [DOI] [PubMed] [Google Scholar]
  • 30.Gestational Hypertension and Preeclampsia: ACOG Practice Bulletin, Number 222. Obstet Gynecol. Jun 2020;135(6):e237–e260. doi: 10.1097/AOG.0000000000003891 [DOI] [PubMed] [Google Scholar]
  • 31.Bernirschke K, Burton G, Baergen R. Pathology of the Human Placenta. 6th ed. Springer-Verlag; 2012. [Google Scholar]
  • 32.Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods. Nov 2009;41(4):1149–60. doi: 10.3758/BRM.41.4.1149 [DOI] [PubMed] [Google Scholar]
  • 33.Team RC. R: A language and environment for statistical computing. https://www.R-project.org
  • 34.Bache S, Wickham H. magittr: A Forward-Pipe Operator for R. https://CRAN.R-project.org/package=magrittr.
  • 35.Champely S. pwr: Basic Functions for Power Analysis. https://CRAN.R-project.org/package=pwr.
  • 36.Heinzen ES J, Atkinson E, Gunderston T, Dougherty G. arsenal: An Arsenal of ‘R’ Functions for Large-Scale Statistical Summaries. https://CRAN.R-project.org/package=arsenal.
  • 37.Müller K. here: A Simpler Way to Find your Files. https://CRAN.R-project.org/package=here.
  • 38.Robinson DH,A., Couch S. broom: Convert Statistical Objects into Tidy Tibbles. https://CRAN.R-project.org/package=broom.
  • 39.Venables WN, Ripley BD Modern Applied Statistics with S. Fourth ed. Springer; 2002. [Google Scholar]
  • 40.Wickham H, Averick M, Bryan J, Chang W, McGowan LD, Francois R. Grolemund G, Hayes A. Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Rsobinson D, Seidel DP, Spinu V, Takahashi K, Vaughn D, Wilke C, Woo K, Yutani H. Welcome to the tidyverse. Journal of Open Source Software. 2019;4(43):1686. doi:doi:10.21105/joss.01686 https://doi.org/10.21105/joss.01686. [Google Scholar]
  • 41.Wickham H, Bryan J. readxl: Read Excel Files. https://CRAN.R-project.org/package=readxl.
  • 42.Zhu H. kableExtra: Construct Complex Table with ‘kable’ and Pipe Syntax. https://CRAN.R-project.org/package=kableExtra.
  • 43.Brosens I, Dixon HG, Robertson WB. Fetal growth retardation and the arteries of the placental bed. Br J Obstet Gynaecol. Sep 1977;84(9):656–63. doi: 10.1111/j.1471-0528.1977.tb12676.x [DOI] [PubMed] [Google Scholar]
  • 44.Khong TY, De Wolf F, Robertson WB, Brosens I. Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-gestational age infants. Br J Obstet Gynaecol. Oct 1986;93(10):1049–59. doi: 10.1111/j.1471-0528.1986.tb07830.x [DOI] [PubMed] [Google Scholar]
  • 45.Lyall F, Robson SC, Bulmer JN. Spiral artery remodeling and trophoblast invasion in preeclampsia and fetal growth restriction: relationship to clinical outcome. Hypertension. Dec 2013;62(6):1046–54. doi: 10.1161/HYPERTENSIONAHA.113.01892 [DOI] [PubMed] [Google Scholar]
  • 46.Ogge G, Chaiworapongsa T, Romero R, et al. Placental lesions associated with maternal underperfusion are more frequent in early-onset than in late-onset preeclampsia. J Perinat Med. Nov 2011;39(6):641–52. doi: 10.1515/jpm.2011.098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Robertson WB, Brosens I, Dixon HG. The pathological response of the vessels of the placental bed to hypertensive pregnancy. J Pathol Bacteriol. Apr 1967;93(2):581–92. doi: 10.1002/path.1700930219 [DOI] [PubMed] [Google Scholar]
  • 48.Kuzmina IY, Hubina-Vakulik GI, Burton GJ. Placental morphometry and Doppler flow velocimetry in cases of chronic human fetal hypoxia. Eur J Obstet Gynecol Reprod Biol. Jun 1 2005;120(2):139–45. doi: 10.1016/j.ejogrb.2004.09.001 [DOI] [PubMed] [Google Scholar]
  • 49.Macara L, Kingdom JC, Kaufmann P, et al. Structural analysis of placental terminal villi from growth-restricted pregnancies with abnormal umbilical artery Doppler waveforms. Placenta. Jan 1996;17(1):37–48. doi: 10.1016/s0143-4004(05)80642-3 [DOI] [PubMed] [Google Scholar]
  • 50.Goldenberg RL, Andrews WW, Faye-Petersen O, Cliver S, Goepfert AR, Hauth JC. The Alabama Preterm Birth Project: placental histology in recurrent spontaneous and indicated preterm birth. Am J Obstet Gynecol. Sep 2006;195(3):792–6. doi: 10.1016/j.ajog.2006.05.050 [DOI] [PubMed] [Google Scholar]
  • 51.Nijman TA, van Vliet EO, Benders MJ, et al. Placental histology in spontaneous and indicated preterm birth: A case control study. Placenta. Dec 2016;48:56–62. doi: 10.1016/j.placenta.2016.10.006 [DOI] [PubMed] [Google Scholar]
  • 52.Raciti P, Sue J, Retamero JA, et al. Clinical Validation of Artificial Intelligence-Augmented Pathology Diagnosis Demonstrates Significant Gains in Diagnostic Accuracy in Prostate Cancer Detection. Arch Pathol Lab Med. Oct 1 2023;147(10):1178–1185. doi: 10.5858/arpa.2022-0066-OA [DOI] [PubMed] [Google Scholar]
  • 53.Rashidi HH, Chen M. Preface: Artificial intelligence (AI), machine learning ML) and digital pathology integration are the next major chapter in our diagnostic pathology and laboratory medicine arena. Semin Diagn Pathol. Mar 2023;40(2):69–70. doi: 10.1053/j.semdp.2023.02.005 [DOI] [PubMed] [Google Scholar]
  • 54.Shafi S, Parwani AV. Artificial intelligence in diagnostic pathology. Diagn Pathol. Oct 3 2023;18(1):109. doi: 10.1186/s13000-023-01375-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Benton SJ, Leavey K, Grynspan D, Cox BJ, Bainbridge SA. The clinical heterogeneity of preeclampsia is related to both placental gene expression and placental histopathology. Am J Obstet Gynecol. Dec 2018;219(6):604 e1–604 e25. doi: 10.1016/j.ajog.2018.09.036 [DOI] [PubMed] [Google Scholar]
  • 56.Jung E, Romero R, Yeo L, et al. The etiology of preeclampsia. Am J Obstet Gynecol. Feb 2022;226(2S):S844–S866. doi: 10.1016/j.ajog.2021.11.1356 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Leavey K, Bainbridge SA, Cox BJ. Large scale aggregate microarray analysis reveals three distinct molecular subclasses of human preeclampsia. PLoS One. 2015;10(2):e0116508. doi: 10.1371/journal.pone.0116508 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Leavey K, Benton SJ, Grynspan D, Kingdom JC, Bainbridge SA, Cox BJ. Unsupervised Placental Gene Expression Profiling Identifies Clinically Relevant Subclasses of Human Preeclampsia. Hypertension. Jul 2016;68(1):137–47. doi: 10.1161/HYPERTENSIONAHA.116.07293 [DOI] [PubMed] [Google Scholar]
  • 59.Roberts JM, Rich-Edwards JW, McElrath TF, Garmire L, Myatt L, Global Pregnancy, C. Subtypes of Preeclampsia: Recognition and Determining Clinical Usefulness. Hypertension. May 5 2021;77(5):1430–1441. doi: 10.1161/HYPERTENSIONAHA.120.14781 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Jaiman S, Romero R, Bhatti G, et al. The role of the placenta in spontaneous preterm labor and delivery with intact membranes. J Perinat Med. Jun 27 2022;50(5):553–566. doi: 10.1515/jpm-2021-0681 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Lien YC, Zhang Z, Cheng Y, et al. Human Placental Transcriptome Reveals Critical Alterations in Inflammation and Energy Metabolism with Fetal Sex Differences in Spontaneous Preterm Birth. Int J MolSci. Jul 23 2021;22(15)doi: 10.3390/ijms22157899 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Morgan TK, Tolosa JE, Mele L, et al. Placental villous hypermaturation is associated with idiopathic preterm birth. J Matern Fetal Neonatal Med. May 2013;26(7):647–53. doi: 10.3109/14767058.2012.746297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Romero R, Dey SK, Fisher SJ. Preterm labor: one syndrome, many causes. Science. Aug 15 2014;345(6198):760–5. doi: 10.1126/science.1251816 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Visser L, van Buggenum H, van der Voorn JP, et al. Maternal vascular malperfusion in spontaneous preterm birth placentas related to clinical outcome of subsequent pregnancy. J Matern Fetal Neonatal Med. Sep 2021;34(17):2759–2764. doi: 10.1080/14767058.2019.1670811 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1971943-supplement-1.pptx (9.1MB, pptx)
2
Download video file (193.4MB, mov)
3
Download video file (45.4MB, mov)
4
NIHMS1971943-supplement-4.pdf (213KB, pdf)

RESOURCES