The human placenta project: Funded studies, imaging technologies, and future directions

Abstract

The placenta plays a critical role in fetal development. It serves as a multi-functional organ that protects and nurtures the fetus during pregnancy. However, despite its importance, the intricacies of placental structure and function in normal and diseased states have remained largely unexplored. Thus, in 2014, the National Institute of Child Health and Human Development launched the Human Placenta Project (HPP). As of May 2023, the HPP has awarded over $101 million in research funds, resulting in 41 funded studies and 459 publications. We conducted a comprehensive review of these studies and publications to identify areas of funded research, advances in those areas, limitations of current research, and continued areas of need. This paper will specifically review the funded studies by the HPP, followed by an in-depth discussion on advances and gaps within placental-focused imaging. We highlight the progress within magnetic reasonance imaging and ultrasound, including development of tools for the assessment of placental function and structure.

1. Introduction

The human placenta functions as a unique multiorgan, simultaneously fulfilling respiratory, renal, hepatic, gastrointestinal, endocrinologic, and immunologic functions. This multifunctional nature modulates both normal pregnancy as well as pathological states that affect both mother and fetus. Consequences of dysfunction include not only immediate sequalae such as insulin resistance, preeclampsia, and growth restriction with concomitant maternal and fetal morbidity, but also longer-term implications for adult neuropsychiatric and cardiovascular health [1-4]. Yet despite its critical role, the human placenta remains one of the least understood organs.

In 2014, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) recognized this lack of understanding and launched the Human Placenta Project (HPP) [5]. Prior to this initiative, existing placental research focused on evaluation at delivery; therefore, the principal goal of the HPP was to better understand human placental structure, development, and function in real time across gestation [6]. Other goals identified included development of noninvasive biomarkers for disease prediction, improved understanding of placental implications for long-term health, development of intervention for abnormal placental development, and improved pregnancy outcomes [6]. To achieve such goals, the NIH has held an annual meeting from 2014 to 2018 and then in 2021 to foster multidisciplinary discussion and review progress to date.

Since the founding of the HPP, the NIH has released several funding opportunities via Requests for Applications (RFAs) and Program Announcements (PAR). The purpose of this article is to summarize the studies and resulting publications funded by the HPP as of May 2023. Our goal was to synthesize published data to identify areas of funded research, advances in those areas, gaps in current published research, and future directions. Given the extensive nature of the HPP, this article will give a broad overview of HPP-funded research categories and then focus more specifically on studies with application in placental imaging.

2. Materials and methods

A review of all studies and resultant publications funded through the HPP was conducted. Eligible studies were identified via the National Institutes of Health (NIH) Research Portfolio Online Reporting Tools (RePORT) on May 3rd, 2023. Studies were considered eligible if they received funding under a known HPP RFA or PAR. These included the following funded announcements: RFA-HD-15-034, RFA-HD-15-030, RFA-HD-15-031, RFA-HD-16-036, RFA-HD-16-037, RFA-HD-17-004, RFA-HD-17-005, RFA-HD-18-003, RFA-HD-18-004, PAR-18-884, and PAR-18-885. The RFA or PAR was used as the search criteria under Funding Opportunity Announcement (FOA) in RePORT. Search results for both studies and publications were extracted separately into Excel using the RePORT export feature.

Studies were first grouped according to research category. These included imaging, specimen collection, and assays performed (omics). Additional data extracted included: source of data (retrospective or prospective, existing data set), participant (e.g. human, non-human primate, mouse, etc.), cohort (uncomplicated pregnancy, pregnancy exposure), and outcomes (e.g. preeclampsia, fetal growth restriction). Further sub-categorization was then performed (Table 1). For imaging studies, these subcategories included magnetic resonance imaging (MRI), ultrasound (US), and use of other imaging modalities. For each study, it was determined if placental function and/or structure were assessed. Placental MRI function studies were subcategorized into technical development, perfusion, oxygenation, diffusion, and metabolism. Placental MRI structure studies were divided among technical development and anatomic assessment. Placental US function studies received the following subcategories: Doppler, technical development, and Power Doppler. Lastly, placental US structure studies were organized as follows: qualitative assessment and quantitative automated analyses.

Table 1.

Study categorization strategy.

Category	Subcategory	Purpose	Imaging Subcategory
Imaging	Magnetic Resonance Imaging (MRI)	placental function	technical development perfusion oxygenation metabolism
		placental structure	technical development anatomic assessment
	Ultrasound (US)	placental function	Doppler technical development Power Doppler
		placental structure	qualitative assessment quantitative automated analyses
	Other
Specimen	Blood/extracellular vesicles Peripheral cells Tissue Urine Amniotic fluid
Omics	Epigenome (e.g DNA, cfDNA) Transcriptome (e.g RNA) Proteome (e.g Mass spectrometry, Antibody-based) Metabolome (e.g Mass spectrometry) Lipidome (e.g Mass spectrometry, profiling) Microbiome

Publications were categorized similarly after determining if the study focus was placental or non-placental. Placental focus was defined as imaging of the placenta in vivo or ex vivo for functional or structural information, use of placental tissue for analysis, and/or correlation of expression with placental tissue. Article type as well as two additional categories were collected. Article type was defined as original research, expert review, systematic review, and case report. The two additional categories were technical development and other to accommodate associated literature of funded studies that were not placental-specific in nature. Technical development encompassed work in other analytic procedures that could be used for placental analysis, particularly on the microscopic to nanoscopic level. Imaging publications that were not placental-focused were excluded from this review. Each original research publication was reviewed by at least two of the authors (C.L.H. and M.J.K.) and assessed for purpose, principle finding, major advance, and identified areas of need. Expert consultation and summative review were then performed (D.M.T., B.F., C.Y.S.).

3. Results

A total of 41 studies were funded by the Human Placenta Project with $101.7 million dollars awarded. There were 459 resultant publications as of May 3rd, 2023 (Table 2). Studies were: imaging only (12, 29%) totaling $32.5 million in funding; imaging and specimen collection with or without planned assay application (11, 27%) with $32 million in funding; or specimen collection with planned assay application (18, 44%) with $37.2 million in funding (Fig. 1). Publications typically focused on either imaging (164, 36%) or specimen collection with assay application (177, 39%) as opposed to combined data together (29, 6%). Slightly less than half of the publications (205, 45%) were specifically focused on the placenta. Those not directly related to the placenta were primarily focused on imaging and analytic techniques to enhance understanding of the placenta or validate techniques that could be applied to the placenta. Examples include MR technical development for metabolism, perfusion, oxygenation, diffusion and segmentation in phantoms and nonpregnant adults and work on acoustofluidics, exosome isolation, DNA or miRNA extraction.

Table 2.

HPP studies, funding, and resultant publications by research category.

Research Category	Number of Studies Funded	Amount Allocated (millions)	Number of Resultant Publications	Placental Publications
Imaging	12 (29%)	$32.5 (32%)	165 (37%)	75 (37%)
Imaging + Specimen	6 (15%)	$18.8 (18%)	23 (5%)	15 (7%)
Imaging + Specimen + Omics	5 (12%)	$13.2 (13%)	8 (2%)	8 (4%)
Specimen + Omics	18 (44%)	$37.2 (37%)	179 (39%)	86 (42%)
Technical Developmenta			43 (9%)	1 (0%)
Other			41 (9%)	20 (10%)
Total	41	$101.7	459	205

^aTechnical Development encompasses work in acoustofluidics, 3D simulation, and other non-omic analytic procedures that could be used for placental analysis, particularly on the microscopic to nanoscopic level.

Fig. 1.

Pie chart representations of funding and resultant publications by research category.

4. Imaging

There were 23 studies and 196 publications that utilized imaging, comprising 56% and 43% of all HPP studies and publications, respectively. Studies were MR only (12, 52%) with $32.4 million in funding; MRI combined with another modality (4, 17%) with $15.8 million in funding; and ultrasound only (7, 30%) with $16.3 million in funding. Publications mirrored this funding application (Table 3). Ninety-eight of the 196 imaging publications (50%) were placental-focused with the majority either MR (53, 54%) or ultrasound alone (25, 26%). Eighty-eight of these publications were original research articles [7-94]; 3 were systematic reviews on ultrasound [95-97]; 4 were expert reviews on MR oxygenation, MR quantitative susceptibility mapping, US-guided photoacoustic imaging of the placenta, and multimodal placental imaging respectively [98-101]; 2 were corrections to previously published articles [102,103]; and 1 was a case report [104].

Table 3.

HPP studies, funding, and resultant publications in imaging. Examples of other imaging include cystoscopy, positron emission tomography, color power angiography, contrast-enhanced and micro computed topography. MRI = magnetic resonance imaging, US = ultrasound.

Imaging	Number of Studies Funded	Amount Allocated (Millions)	Number of Resultant Publications	Placental Publications	Placental Original Research
MRI	12 (52%)	$ 32.4 (50%)	119 (61%)	53 (54%)	49
MRI + US	2 (9%)	$ 8.3 (13%)	9 (5%)	7 (7%)	6
MRI + US +other	1 (4%)	$ 4.3 (7%)	2 (1%)	1 (1%)	0
MRI + other	1 (4%)	$ 3.2 (5%)	7 (4%)	3 (3%)	3
US	7 (30%)	$ 16.3 (25%)	43 (22%)	25 (26%)	22
US + other			4 (2%)	3 (3%)	3
Other			12 (6%)	6 (6%)	5
Total	23	$64.5	196	98	88

5. Magnetic resonance imaging

There were 58 original research articles that utilized MRI [7-64], all but six [59-64] of which characterized aspects of placental function and 34 aspects of placental structure [7-18,29-39,47-51,59-64] (Table 4). Within placental function, 19 studies covered aspects of technical development [7-25], 15 placental perfusion [26-40], 19 placental oxygenation [12,36-53], 5 placental diffusion [34,41,47,51,54], and 4 placental metabolism [55-58]. Within placental structure, 8 studies focused on aspects of technical development [13-18,59-61] and 28 anatomic assessment [7-13,29-39,47-51,62-64].

Table 4.

Original research publications in imaging by subcategory. 14 studies included more than one modality: 7 magnetic resonance imaging (MRI) and ultrasound (US), 3 US and other, and 4 MRI and other.

Category	Modality	Purpose	Subcategory
Imaging (88)	Magnetic Resonance Imaging (58)	placental function (52)	technical development (19) perfusion (15) oxygenation (19) diffusion (5) metabolism (4)
		placental structure (34)	technical development (9) anatomic assessment (26)
	Ultrasound (32)	placental function (24)	Doppler (17) technical development (2) power Doppler (6)
		placental structure (13)	qualitative assessment (5) quantitative automated analyses (9)
	Other (12)

6. MR advances

6.1. Placental function

6.1.1. Technical development

MR technical development from the HPP has worked with animals, retrospective data sets, and prospective recruitment of pregnant people to assist in the development of novel techniques for assessment of placental function. In animals, work has been done in pregnant mice to characterize angiogenesis at implantation [7], biotin transporter activity [8], and placental blood volumes [23]. Feasibility and proof-of-principle studies of ferumoxytol-enhanced MRI at the maternal-fetal interface were found to have minimal effects on placenta or fetus [9], correlate oxygenation with spiral arteries [10], and evaluate placental inflammation in pregnant rhesus macques [11,25].

Human characterization worked toward assessment with uterine activity during pregnancy [14], motion correction [15-17,24], placental flattening [18], and automated segmentation of the placenta to obtain functional imaging data [19]. Other studies investigated prospective in vivo assessment of placental dysfunction via hemoglobin dissociation curves [20] and blood hemodynamics [12]. Assessment of in-vivo fetal-blood oximetry was accomplished by utilizing a calibration equation obtained from T₂ and HbO2 values of ex-vivo fetal blood from umbilical cord blood samples [21]. Further human feasibility data has included assessment of diffusion using intravoxel incoherent motion (IVIM) [12,13] and ex-vivo assessment of perfusion of the umbilical artery and intervillous space [22].

6.1.2. Perfusion

Multiple HPP investigators have worked to develop in vivo animal and human models for assessment of placental perfusion. In animals, these include application of dynamic contrast-enhanced (DCE-MRI) with gadoteridol [28,29,37-40] and ferumoxytol [31,35], arterial spin labeling (ASL) via flow-sensitive alternating inversion recovery (FAIR) [31], and 4D flow MRI [32] for assessment of perfusion in pregnant macaques. DCE-MRI was found to demonstrate significantly reduced perfusion in pregnant macaques with fetal growth restriction (FGR) [38], chronic first trimester alcohol exposure [37,40], and with pathology due to Zika virus [35,39]. In addition, DCE-MRI demonstrated significantly reduced perfusion in pregnant macaques on a chronic western-style diet, but this reduction could be ameliorated with diet reversal [28]. DCE-MRI also demonstrated reduced perfusion with aberrant uterine contractility in pregnant rats [36] and mouse models of preeclampsia and growth-restriction [34]. Based on current modeling [28,29,37-39] and the large molecular size of ferumoxytoxl [31,35], DCE-MRI primarily reflects the maternal side of placental blood flow.

Various methods have been explored for the assessment of perfusion measurement in uncomplicated human pregnancy. These include a free-breathing pseudocontinous (pCASL) ASL technique [26] and a velocity-selective method (VS-ASL) [27]. In the setting of ischemic placental disease (gestational hypertension, preeclampsia, and FGR), pCASL-based perfusion was found to significantly decrease when measured at 16 weeks [30,33]. pCASL reflects maternal-side placental blood flow, while VS-ASL reflects both maternal and fetal blood flow.

6.1.3. Oxygenation

Similar to perfusion, animal and human models of placental oxygenation have also been developed by HPP investigators. Overlap can be seen among groups assessing perfusion and oxygenation metrics [36-39,47]. Oxygenation has been assessed using blood oxygen level-dependent (BOLD) imaging via relaxometry with T₂* and R₂* (1/T₂*) mapping [36-39,41,43,46-52]. Others used BOLD to assess oxygen kinetics and have assessed time to plateau (TTP) [42,44]. Oxygenation has also been measured through quantitative susceptibility mapping (QSM) [12,45]. Combined T₁ and T₂-mapping has also been explored for representation of oxygenation, as T₂ is not impacted by local magnetic field inhomogeneities like T₂* [53].

Similar to effects on perfusion, reduced oxygenation was also observed in pregnant macaques with FGR [38] and chronic first trimester alcohol exposure [37,40] as well as with aberrant uterine contractility in pregnant rats [36]. In Zika-infected pregnant macaques, higher-than-expected oxygenation of maternal blood was found within the placenta, which the authors suggested may reflect decreased oxygen permeability or decreased fetal demand in affected placentas [39].

In humans, global placental T₂* has been shown to decrease over gestation. Recently T₂* has been further described as decreasing continuously from a high plateau level early in gestation, through an inflection point at 30 weeks, and approaching a second, lower plateau in late gestation, based off a cohort of 316 pregnancies [50]. In a study that included anterior and fundal placentas, there was a decline in T₂* from center to periphery in each individual placental lobule, thought to represent oxygen exchange between maternal and fetal circulation within the intervillous space [47]. Regarding clinical correlation, T₂*is significantly reduced in placentas with adverse outcomes [48,50], maternal chronic hypertension [41], and congenital heart disease [52] compared to unaffected placentas. Variation in R₂* has been suggested based on placental location, with lower values in anterior compared to posterior placentas derived from 3D volumetric analysis [43]. Maternal position and contractions have also been found to cause variation, with global and regional differences in T₂*/R₂* and regional time to plateau (TTP) [44,46,49]. Moreover, decreases in placental T₂* with contractions have been found to correlate with decreased in T₂* in the fetal liver and brain [46]. In addition, with maternal hypoxemia, longer TTP has been observed in the smaller of discordant monozygotic twin pairs [42].

6.1.4. Diffusion

Diffusion-weighted (DW) MRI has been investigated for its complementary information, often used as part of a combined approach for functional data. Diffusion-weighted imaging (DWI) is sensitive to water motion which may include intra- and extra-cellular water diffusion and may be a marker for early perfusion while also reflecting in vivo morphology. In mouse models of preeclampsia and growth-restriction, similar to perfusion, diffusion was also decreased in these diseased states [34]. In human pregnancy, examples included combined relaxometry-diffusion using T₂* and apparent diffusion coefficient (ADC) mapping [47,51] and T₁ and ADC mapping, with an acquisition time under 2 min [54]. These studies describe trends for decreases in T₂* and diffusivity with increasing gestational age [47] and placental disease (preeclampsia, fetal growth restriction) [51,54]. Another combined MR examination (T2-weighted imaging, T₂*, T₁ and ADC mapping) recently demonstrated varying placental phenotypes in pregnant individuals with chronic hypertension; however, values overlapped with the control group of healthy pregnant individuals [41].

6.1.5. Metabolism

Work on placental metabolism within MR imaging is in its early stages. Preliminary work has been applied to look at hyperpolarized ¹³C MRI in animals for the metabolism of substrates such as pyruvate, lactate, and urea in the placenta [55-57]. Luo et al. demonstrated feasibility of glucose chemical saturation exchange transfer (CEST) in humans for non-invasive measurement of glucose transport in the placenta [58].

6.2. Placental structure

6.2.1. Technical development

Efforts to better delineate placental structure included some animal and many human pregnancies. Liposomal gadolinium was used in a pregnant mouse model to specify the retroplacental clear space as well as morphologic volume changes thorough gestation [59,61]. Algorithms for placental flattening [18,60] and motion correction [15,17] were used for structural assessment as well, with development of placental volume reconstruction using a dataset of human fetuses with minor and severe motion [16] and characterization of uterine activity during pregnancy [14]. Ansiotropic IVIM models were applied for assessment of human placental microvascular structure and tissue microstructure [13].

6.2.2. Anatomic assessment

Most studies relied on manual regions of interest (ROIs) for placental identification. Similarly, division of the placenta into anatomic units has occurred through manual and post-processing techniques. Manual ROIs with DCE-MRI have been applied for placental lobe and cotyledon specification [10-12,29,34,35,37-39], volume measurements [29,35], maternal and fetal tissue delineation [9], implantation site identification [7], and for delineation of uterine vasculature as well as the intervillous space [31,32]. Others have proposed dividing the placenta into various compartments via mathematical modeling [8]. In human placentas, manual segmentation has been used for volume quantification [30,33, 35,47-50] and anatomic division [12,13,36,47,51,64]. Recently, automated placental segmentation has begun and has been applied to placental volumes [62].

7. MR gaps

When considering the gaps remaining after advances from the HPP, it is useful to frame current knowledge in the context of Standards for the Reporting of Diagnostic Accuracy Studies (STARD) [105]. Diagnostic testing is most useful when a medical test is available that provides useful information on the health of the patient, a target condition is defined, there is an accepted clinical reference standard, the test has sufficient sensitivity and specificity, and testing results in a clear intention for use (diagnosis, screening, prediction, prognosis).

On this basis there are several current limitations in MR knowledge. Medical testing is available, but there is not yet an agreed upon best approach and much is needed in the manner of technical development prior to pursuit of diagnostic testing. For MR-based techniques challenges include the need for correction or adjustment due to maternal breathing, fetal motion, contractions, maternal position, placental position. Variation in maternal blood volume, blood flow, blood oxygenation, and blood glucose levels pose difficulty with functional data assessment. The optimal MR approaches for perfusion, oxygenation, and diffusion are not yet clear. There is also variation among 2D versus 3D measurement ascertainment and in determination of ROIs [19,53]. In addition, most ROIs previously were manually segmented. Movement towards automatic segmentation can remove inherent subjectivity in measurement [19,62]. Finally, current published studies primarily assessed global measurements, though evidence suggests that there is likely regional anatomic variation in placental function [22,33,46]. Whether the measurement needs to be cotyledon-specific, maternal--sided, fetal-sided, based out of the intervillous space, or a composite of these areas also remains to be determined. Due to these issues, further optimization and standardization of MR technique is needed to enable comparison of the literature and prior to clinical application for diagnostic testing.

There is also a need to establish a clinical reference standard for MR functional imaging, specifically gestationalage adjusted normal ranges [40,41,48]. Some headway has been made for T₂* [50]. Larger sample size of uncomplicated human pregnancies are still needed, as small sample size is a limitation of much of the current published data [12,17, 21,26,27,43,45,49,51,52]. In addition, larger cohorts of disease states are also required, as the composites of adverse outcomes are not sufficiently predictive at present [50]. As these sample sizes grow, further work will elucidate if sufficient sensitivity and specificity can be found to aid prediction of and ultimately intervention for placental-mediated disease.

8. Ultrasound

There were 32 original research publications that utilized ultrasound [28-30,37-40, 65-89]: 24 covered aspects of placental function [28-30, 37-40, 65-81] and 13 placental structure [28,39,79-89]. Within placental function, 17 focused on Doppler [28-30,37-40,65-74], 3 on technical development [28,75], and 6 on Power Doppler [76-81]. For placental structure, 4 has components of general assessment [28,39,79, 80,82] and 9 ultrasound studies worked on automation [80,81,83-89].

9. US advances

9.1. Placental function

9.1.1. Doppler

Doppler has been used by many for attempted quantification of uteroplacental blood flow with limited association to clinical outcomes. Most commonly, Doppler has been used to assess uterine artery pulsatility index as a marker for impedance of uteroplacental circulation with mixed success [28,30]. Uterine artery pulsatility index does vary by fetal sex [71]. Another common method described has included calculated blood flow of both the uterine artery (maternal vessel, termed cQ_UA) and umbilical vein (fetal vessel, termed cQ_UV) [28,29,37,38,40]; however, measurements limited to these two major vessels seemingly lack sensitivity to detect placental dysfunction, likely as they are not completely reflective of the intervillous space [39,66].

Others have studied umbilical artery (UA) velocity wave reflections as a potential marker of dysfunction, hypothesizing that elevated placental vascular resistance may be observed as an elevated pulse wave from the fetal heart [65,70] Initial work was undertaken in mouse models [65,67] and found wave reflections more sensitive to changes in placental vascular resistance than UA pulsatility index [68] and correlated negative flow reflected waveforms with increased terminal impedance in the placenta [74]. Subsequent studies transitioned to human pregnancy [69,70], and have found significantly elevated wave reflections in pathologic outcomes of maternal vascular malperfusion and fetal vascular malperfusion [72].

Superb microvascular imaging (SMI) has been paired with color Doppler to determine resistive indices of the spiral arteries and intervillous space, which were significantly more elevated in the late third trimester for preeclampsia and fetal growth restriction [73]. However, the value added of this finding is limited, given the association is only with late gestational age and it has not yet been validated by larger studies.

9.1.2. Technical development

Work toward functional assessment in ultrasound included modeling of flow from the spiral arteries into the intervillous space throughout gestation based on pulsed wave Doppler ultrasound in uncomplicated pregnancies at 11-13 weeks [75]. The model found that villous structure appears under the influence of spiral artery blood flow, as free spaces must develop distal to spiral arteries to allow blood flow to penetrate among villous tree branching.

Contrast-enhanced ultrasound with microbubbles has been used in pregnant macaques to visualize individual spiral arteries and determine spiral artery transit time [28]. Transit time was found to be increased in those exposed to a chronic Western-style diet, correlating with decreased placental perfusion observed on MR imaging and decreased uterine arteiy blood flow as determined by Doppler [28].

9.1.3. Power Doppler

Power Doppler (PD) can used as a measure of vascularity in tissue. 3-dimensional PD (3D PD) has been applied in the first trimester (11-13 weeks) in human pregnancy for assessment of placental vascularity. Published assessments include a diagnostic marker for placenta accreta spectrum along the uteroplacental interface, termed the largest area of confluent 3D PD signal (A_con). In PAS, A_con is quantitatively increased, in keeping with known neovascularization along the bladder-serosal interface [76]. This measurement was found to be best achieved with a filled bladder as opposed to unfilled [77]. Another PD-derived metric is 3D fractional moving blood volume (3D-FMBV), a validated method for estimating tissue perfusion using ultrasound [80]. Applied at the uteroplacental interface, pregnancies that develop preeclampsia have reduced 3D-FMBV whereas pregnancies with small for gestational age (SGA) pregnancies resulted in similar vascularity to normotensive appropriate for gestational age (AGA) pregnancies [80]. This technique was recently automated [81]. Also using automated 3D PD, spiral artery jet number, area, and PD area close to the uteroplacental interface (2-3 mm) was found to be smaller in preeclampsia and some pregnant individuals with SGA pregnancies as compared to those with normotensive AGA pregnancies [79].

In contrast to 3D approaches, others have used a 2D approach and demonstrated that placental vascular pulsatility varies throughout the maternal cardiac cycle [78]. As pulsatility can reflect vascular modeling and resistance, this method holds promise as a potential marker for abnormal spiral artery remodeling that enables inter-subject comparison, as patients can serve as their own controls. Such insight is important as this cycle variation is not yet well accounted for in 3D techniques.

9.2. Placental structure

9.2.1. Qualitative assessment

Some features of placental structure were also assessed in US studies. In animals, contrast-enhanced ultrasound has been used for spiral artery identification [28,39]. Similar to vascularity, 3D-derived placental volume has been used in the assessment of the first trimester placenta in human pregnancy. Placentas of those who go on to develop pre-eclampsia and small for gestational age pregnancies had smaller placentas than normotensive AGA pregnancies [80]. Number of spiral arteries and size (cross-sectional area) have also been determined using 3D PD ultrasound data [79]. Ultrasound has also been used to aid in the application of optical spectroscopy (see Other Imaging) for delineation of placental oxygenation to provide tissue layer morphology (e.g. adipose, rectus/uterus, and placental tissue) and physiologic properties [82],

9.2.2. Quantitative automated analyses

A move toward automated tools for standardized structural evaluation was observed in many of the ultrasound studies. Described in the literature are algorithms for semi-automated 3D placental volume assessment [80,88], structural detection of the placenta, myometrium, and subcutaneous tissue [83], automated lacunae localization in PAS [84], knowledge-guided pretext learning for the uteroplacental interface in PAS [85], and now fully automated placental volumes using deep learning [81,86,87,89]. The value added in prediction for small for gestational age by first trimester volume is potentially promising. While the area under the curve for first trimester volumes has ranged from 0.61 to 0.78,^86,89 this improved significantly to 0.897 when added to additional clinical modeling [89].

10. US gaps

Similar limitations as outlined for MR exist within US data. The ideal US test for placental disease is unknown and there remain several technical limitations. There is a need to standardize US machine settings, determine the best method for measurement and of which space (maternal, fetal, and intervillous), and automate measurement ascertainment. Reduction of operator dependency will be necessary prior to larger clinical trials given limited reproducibility or replication. Additional consideration for 3D PD includes the pulsatile nature of the vascular bed [78]. Techniques must also be developed that account for variation in placental location [78] and impact of fetal sex given recently discovered differences [71].

For development of diagnostic testing, larger sample sizes will be needed to create a reference standard that is distinguishable from diseases states, as most studies are limited to small sample sizes [72,73,81, 83,86] and/or in uncomplicated human pregnancy [67,68,81]. Larger and newer data sets are need for automation [87,89] given the rapid evolution of technology and for disease association.

11. Other imaging

There were 12 studies that utilized other imaging techniques aside from MR and US for evaluation of the placenta [8,23,61,63,67,74,82, 90-94]. X-ray micro-computed tomography (CT) [67,74,90-92] has been used for elucidation of placenta vasculature in animal models. A combination of micro-CT and scanning electron microscopy has been used to assess placental adaptations to chronic fetal hypoxia in mice [91]. Pre-operative cystoscopy for the assessment of PAS severity when placentation is over the prior scar and in proximity to the bladder wall has also been investigated [93].

Often additional imaging techniques were used to complement or validate MR findings. For example, in pregnant mice, contrast-enhanced CT has been used to validate placental fractional blood volume [23], retroplacental clear space [61], and placental margins [63] as determined by contrast-enhanced MR with liposomal gadolinium, and positron emission topography was used to correlate with biotin transporter activity [8]. Use of such imaging in human pregnancy remains limited to due ionizing radiation exposure.

An evolving field for placenta evaluation is optical imaging, with many preclinical investigations reported. More recent clinical applications have included optic spectroscopy [82] coupled with ultrasound imaging and photoacoustic imaging [94] for monitoring of placental oxygenation.

12. Limitations

We herein report on all placental-focused imaging publications resulting from HPP funding, though we recognize that this does not encompass all placenta research and advancement. Pre-clinical studies not directly applied to the placenta are not included. All publications were reviewed based on the NIH Reporter system. It is possible that manuscripts not yet associated within this system but published are not included in this review.

13. Future directions

The HPP has enhanced our real-time knowledge of the placenta, but numerous gaps remain. Next steps include determining the most effective non-invasive methods for assessment of placental function. The technical changes faced by current imaging methods must be addressed, including determining the appropriate regions of interest for measurement. Translation of MR and US to clinical application for diagnostic testing at present is likely hindered by the inter-observer variability and different institutional protocols. Therefore, working toward reproducible standardized ascertainment, automated analyses, and reporting is paramount. This includes creating an environment that fosters encouragement of shared algorithms, machine software, calculations, and methods. More systemic issues must also be acknowledged that limit current innovation – profit-driven obstacles at the manufacturer level including machine-specific software and current and historical under-funding for women’s health research [106].

As processes become better defined, an additional step will be to reliably and reproducibly correlate imaging metrics with normal placental function throughout gestation and with placental disease pathology. Only then can clear associations with disease be solidified and the ultimate goal of understanding the placental disease achieved – a clinically useful tool for disease prediction that enables intervention for such pathology.

Abbreviation guide

HPP

Human Placenta Project

NICHD

National Institute of Child Health and Human Development

FGR

Fetal growth restriction

MRI

Magnetic resonance imaging

US

Ultrasound

DCE-MRI

Dynamic contrast enhanced MRI – MR acquisition obtained during the passage of a contrast agent

ASL

Arterial spin labeling: noninvasive perfusion technique in which endogenous water in labeled upon a selected plane via radiofrequency and magnetic field gradient pulses

pCASL

Pseudocontinuous ASL: ASL method with enhanced signal and reduced signal-to-noise ratio compared to pulsed labeling

VSASL

Velocity-selective ASL: noninvasive perfusion technique in which blood is labeled based on velocity

FAIR

Flow-sensitive alternating inversion recovery: noninvasive ASL perfusion technique in which pulse sequences generate tagged and control images

BOLD-MRI

Blood oxygen level-dependent MRI: noninvasive technique that when applied in the placenta reflects oxygenation based on changes in the ratio of oxyhemoglobin to deoxyhemoglobin

QSM

Quantitative Susceptibility Mapping: MRI measurements of the magnetic susceptibility of a tissue. For oxygenation, it is directly related to the hemoglobin saturation

ROI

Region of interest

T₂*

Time constant for the decay of transverse magnetization caused by a combination of spin-spin relaxation and magnetic field inhomogeneity. For oxygenation, based on changes between paramagnetic deoxyhemoglobin and non-paramagnetic oxyhemoglobin

R₂*

Apparent proton relaxation rate, equivalent to 1/T₂*

TTP

Time to Plateau: time that the signal change due to regional oxygenation reaches to the plateau

DWI

Diffusion-weighted imaging: MRI acquisition based on measuring the random Brownian motion of water protons within a voxel of tissue

IVIM

Intravoxel incoherent motion: characterizing the microscopic motions of water in each image voxel, using diffusion-weighted imaging

ADC

Apparent diffusion coefficient: measure of the magnitude of diffusion of water protons within tissue, calculated using diffusion-weighted imaging

CEST

Chemical exchange saturation transfer: MRI contrast approach in which endogenous or exogenous compounds with exchangeable protons are selectively saturated and, after transfer of this saturation, detected through an attenuated water signal

3D-FMBV

3-dimensional fractional moving blood volume