Abstract
There’s a paucity of robust normal fractional limb and organ volume standards from a large and diverse ethnic population. The Fetal 3D Study was designed to develop research and clinical applications for fetal soft tissue and organ volume assessment. The NICHD Fetal Growth Studies (2009–2013) collected 2D and 3D fetal volumes. In the Fetal 3D Study (2015–2019), sonographers performed longitudinal 2D and 3D measurements for specific fetal anatomical structures in research ultrasounds of singletons and dichorionic twins. The primary aim was to establish standards for fetal body composition and organ volumes, overall and by maternal race/ethnicity, and determine whether these standards vary for twins versus singletons. We describe the study design, methods, and details about reviewer training. Basic characteristics of this cohort, with their corresponding distributions of fetal 3D measurements by anatomical structure, are summarized. This investigation is responsive to critical data gaps in understanding serial changes in fetal subcutaneous fat, lean body mass, and organ volume in association with pregnancy complications. In the future, this cohort can answer critical questions regarding the potential influence of maternal characteristics, lifestyle factors, nutrition, and biomarker and chemical data on longitudinal measures of fetal subcutaneous fat, lean body mass, and organ volumes.
Keywords: 3D ultrasound, fetal body composition, fetal growth, fetal volume
Abbreviations
- ASCTT
abdominal subcutaneous tissue thickness
- BMI
body mass index
- CV
coefficient of variation
- EFW
estimated fetal weight
- NICHD
National Institute of Child Health and Human Development
- SCTT
subcutaneous tissue thickness
Fetal growth is an important predictor of health and disease under the developmental origins of health and disease (DOHaD) paradigm (1–4). Despite this knowledge, precise classification of abnormal fetal growth as “restriction” or “overgrowth” during pregnancy remains an important clinical challenge. Conventional 2D ultrasonography emphasizes calculating an estimated fetal weight (EFW) based on fetal head, abdomen, and femur measurements because formulas based on these have been found to have the lowest mean absolute error in predicting birth weight (5). A comparison of accuracy of formulas found that formulas based on these measures generally performed similarly. They have an 80%–87% accuracy in predicting birth weight within 15%; however, prediction is more inaccurate at the extremes of birth weight, which can have important clinical consequences (5–7). Undetected fetal growth restriction is associated with increased neonatal morbidity (8). At the other extreme, error in predicting macrosomia can have indications for delivery planning (9). Specifically, cases incorrectly predicted to have small fetuses may be allowed to deliver vaginally and be at risk for shoulder dystocia when the fetus is actually macrosomic, while cases incorrectly predicted to have macrosomic fetuses may be unnecessarily subjected to cesarean delivery.
In addition to standard measurement error, inaccuracy in estimating fetal size has been attributed in part to using 2D measures to approximate 3D fetal volume. Calculations for EFW primarily emphasize skeletal measurements and not soft tissue development. Weight is a measure of mass (size × density), and density is not considered when using fetal size based on 2D measures (10). Lean body mass and fat mass have different densities, and differences in fetal body composition have been found to explain a substantial amount of variance in normal birth weight (11). Given these limitations of calculating EFW with traditional formulas, studies have investigated incorporating 2D measures of fetal body composition, defined as evaluation of different body compartments including body fat and lean tissue (muscle and bone) (12). An EFW formula with mid-thigh soft tissue thickness, and another with visceral adipose tissue and abdominal subcutaneous fat thickness, had modest reductions in the error between EFW and birth weight over that of traditional EFW formulas, as well as an EFW formula with fetal cheek-to-cheek diameter among macrosomic (>4,000 g) fetuses (13–16). Incorporating fetal fractional limb volumes from 3D ultrasound also holds promise to improve accuracy in fetal weight estimation or detection of late fetal growth restriction (17–22). Fractional limb volume (FLV) measures are 3D soft tissue parameters that are derived from mid-portions of the arm or thigh as proxies for fetal nutritional status (18). However, one of the major barriers to clinical acceptance has been the paucity of robust normal fractional limb volume standards from a larger and more diverse ethnic population. The Fetal 3D Study can overcome these barriers.
In addition to calculating fractional limb volumes, 3D ultrasonography now permits more robust characterization of fetal body composition in ways that are not possible using conventional 2D (12). Body composition may be important in line with the developmental-origins-of-health-and-disease paradigm since preliminary work suggests that neonatal percent body fat may have independent implications for future risk of metabolic syndrome over that of birth weight alone (1, 12, 23, 24). Fetal body composition may become altered in pregnancies complicated by fetal growth restriction as a result of chronic metabolic impairment (25, 26). Soft tissue parameters may also represent imaging biomarkers that indirectly reflect fetal nutritional status as a physiologic adaptation to altered energy demands in pregnancy conditions such as gestational diabetes (19, 27–30). However, serial longitudinal changes in fetal body composition in association with other pregnancy complications such as preeclampsia as indicators of fetal morbidity and maternal disease are unknown. Expanding assessment of fetal body composition using 3D may improve detection of changes over that of conventional 2D imaging methods, which could lead to better prediction and management of pregnancies where the infant is small or large for gestational age.
In addition to body composition, organ sizes such as the kidney and liver volumes may be altered with pregnancy conditions such as the infant being small for gestational age and gestational diabetes (31–34). Although there have been studies published regarding reference ranges for organ sizes, most of these had small numbers, many were cross-sectional, and the times of measurements were not standardized (35–43). Moreover, little is known of the etiology of daily lifestyle exposures in healthy pregnancies, such as diet, physical activity, air pollution, sleep, etc., on fetal body composition and organ volumes, or how maternal factors such as body mass index (BMI) may influence fetal volumes. Knowledge of these associations could help increase our understanding of the etiology of fetal and subsequent neonatal morbidity, and guide future interventions such as improved maternal disease control, individualized pregnancy monitoring and determining timing of delivery.
The Fetal 3D Study was designed to answer these critical data gaps. Data collected in the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Fetal Growth Studies included longitudinal acquisition of 2D fetal biometrics and 3D fetal volumes, which were banked in singleton and dichorionic twin gestations. The primary aim of the Fetal 3D Study was to establish standards for fetal body composition and organ volumes by maternal race/ethnicity, and determine whether they vary for dichorionic twins compared with singletons (for the volume types that were collected). Our original intent was to follow the NICHD Fetal Growth Studies—Singletons procedure that developed racial- and ethnic-specific fetal growth standards for 2D fetal biometrics and EFW because fetal size varies by ethnicity and country (44–46). Also, our group found that a unified fetal growth standard differentially classified small or large for gestational age for different groups (47). In a recent study by Ikenoue et al. (48) in Japan, fetal fractional limb volumes were smaller across pregnancy than that previously published in a US cohort, with differences becoming more pronounced with advancing gestation. These findings suggest that 3D differences in fetal growth also may vary by ethnicity and country. Since that time, a multiethnic, unified NICHD fetal growth standard has also been developed given recognition that inclusion of self-reported race and ethnicity in clinical algorithms may create unintended consequences for diagnosis and intervention (49–52). Unified standards have also been developed by 2 other international fetal growth studies (46, 53, 54). Therefore, we plan to create both a multiethnic, unified standard for fetal body composition and organ volumes and standards according to race and ethnicity. Secondary aims included:
Assess the relationship between gravid diseases and longitudinal changes in fetal body composition and organ volumes, and compare with those fetuses growing under optimal circumstances (i.e., singleton gestations included in the development of the fetal growth standard (44, 52, 55))
Investigate organ volumes (and ratios to fetal weight) and mass (fat) size in association with gravid or neonatal complications (e.g., kidney to fetal weight in fetal growth restriction)
Determine whether factors including maternal BMI, weight gain, longitudinal changes in maternal body composition, nutrition, and lifestyle factors would modify the relationship between longitudinal changes in fetal body and organ volumes and gravid diseases, and whether they vary by plurality (singleton or twin)
Explore the association of biomarkers (including cardiometabolic and nutritional) with longitudinal changes in fetal body composition and organ volumes
Investigate chemical environmental exposures in association with fetal body composition and organ volumes
METHODS
Participants
Two-dimensional (2D) and 3D imaging data were collected from women enrolled in the NICHD Fetal Growth Studies—Singletons and Dichorionic Twins at 12 and 8 US sites, respectively, from 2009–2013, including 2,334 low-risk pregnant women with BMIs (weight (kg)/ height (m)2) of 19–29 from 4 self-identified race and ethnic groups (Asian or Pacific Islander, Hispanic, non-Hispanic Black, and non-Hispanic White), 468 women with BMIs 30–45, and 171 women with dichorionic twin gestations from 2009–2013 (56–58). The Fetal 3D Study included all women with banked imaging data for measurement of 2D abdominal body wall thickness and 3D fetal body composition and volumes that occurred between September 2015 and September 2019. (Clinicaltrials.gov identifiers: NCT00912132, NCT03266198.)
Ultrasound data collection
In the NICHD Fetal Growth Studies—Singletons and Dichorionic Twins, women were recruited in the first trimester of pregnancy and followed through pregnancy to delivery (56). A requirement to participate in the study was accurate dating: The ultrasound estimate of gestation was between 8 weeks 0 days and 13 weeks 6 days and matched the last menstrual period–based gestational age within 5 days for women between 8 weeks 0 days and 10 weeks 6 days; 6 days for those between 11 weeks 0 days and 12 weeks 6 days; and 7 days for participants between 13 weeks 0 days and 13 weeks 6 days. All women gave informed consent, and institutional review board approval was obtained at all participating sites and the data coordinating centers. Details on recruitment, retention, and study procedures for the singletons and twins have previously been published (44, 56–58). In brief, following an initial gestational dating sonogram at 10–13 weeks of gestation, each woman with a singleton was randomized to one of 4 follow-up visit schedules with 5 additional sonograms for fetal biometry plus additional image and 3D volume acquisition for later analysis as outlined in Table 1. Twins were randomized to follow up at one of 2 schedules (16, 20, 24, 28, 32, and 35 or 18, 22, 26, 30, 34, and 36 gestational weeks). Research visits were allowed within 1 week to allow for flexibility in scheduling.
Table 1.
Randomization Schedulea for Singletons, National Institute of Child Health and Human Development Fetal Growth Studies, United States, 2009–2013
| Group | Visit 1 | Visit 2 | Visit 3 | Visit 4 | Visit 5 |
|---|---|---|---|---|---|
| A | 16 | 24 | 30 | 34 | 38 |
| B | 18 | 26 | 31 | 35 | 39 |
| C | 20 | 28 | 32 | 36 | 40 |
| D | 22 | 29 | 33 | 37 | 41 |
a Research visits were allowed within 1 week to allow for flexibility in scheduling.
Sonographers underwent didactic and hands-on training, were credentialed prior to study start, and followed a standardized protocol for ultrasound acquisition. Scans were performed using Voluson E8 (GE Healthcare, Milwaukee, Wisconsin). Blank images of the 2D abdomen were collected as part of the quality assurance for development of the fetal growth standard and therefore were available for measurement of subcutaneous tissue thickness (SCTT) (59). In addition to traditional 2D biometrics, the following volumes were targeted for collection among singletons: in the first trimester, fetus and gestational sac; in the second and third trimesters, cerebellum, chest, heart, abdomen, pelvis, arm, and thigh. In twins, the volumes collected were, in the first trimester, fetus and gestational sac, and in the second and third trimesters, thigh (both twins). The protocol outlined collection of 2 volumes for each anatomical structure; however, if time was limited, then to prioritize collection of the cerebellum, abdomen, and thigh volumes. All imaging data were algorithmically or manually reviewed by the data coordinating centers to check for and remove any images with personal identifying information. Details of the 3D volume acquisition are outlined in Web Appendix 1 (available at https://doi.org/10.1093/aje/kwad210).
Fetal 3D cataloguing and measurement
The sonology team included 6 physicians with prenatal diagnostic ultrasound expertise and 10 registered diagnostic medical sonographers. All available scans were reviewed for specific 2D and 3D images relevant to this study. Reviewers performed cataloguing to identify what was contained within the scan for each intended anatomical structure, all information pertaining to scan identification, image and/or volume identification, and grading of quality. Scans were accessed through the data coordinating centers by a secure logon to an electronic data capture system. Once logged into the system, the reviewer downloaded from cloud storage the files in their queue and imported them into Viewpoint 6 (GE Healthcare, Milwaukee, Wisconsin) software.
After the catalog was complete, W.L. and L.M. assigned a pair of expert reviewers from the sonology team according to specific anatomical structures as listed in Table 2. The experts reviewed images of the anatomical structure to decide how to best measure parameters within the anatomical structure (for example, cerebellar volume and transcerebellar diameter within the fetal head) and established procedures based on the published literature and experience. They then followed a protocol for standardization. See Web Appendix 2 for details. The teams of expert reviewers assigned for each organ system then trained the rest of the reviewers. Remaining cases were divided among reviewers with all ultrasound scans belonging to each woman being assigned to a single reviewer. Serial measurement of available 3D fractional arm and thigh volumes, arm and thigh body composition, abdominal wall body composition, and visceral volumes were performed in 2D and 3D banked images for all visits. Total as well as lean fractional limb volumes were measured as per Figure 1. Fractional limb fat volumes can be calculated by subtracting the lean from the total volume measurements. In twin gestations, both twins were measured. A list of the parameters and biometric measurements are in Table 2. Details of measurement procedures are also presented in Web Appendix 3.
Table 2.
Summary of Measurements, National Institute of Child Health and Human Development Fetal 3D Study, United States, 2015–2019
| Anatomical Structure and Parameter | Units |
|---|---|
| Arm | |
| Humerus diaphysis length (HDL) | cm |
| Mid arm circumference (arm C) | cm |
| Mid-arm area (MAA) | cm2 |
| Mid-arm lean area (MALA) | cm2 |
| Mid lean arm circumference (mid lean arm C) | cm |
| Maximum arm subcutaneous tissue thickness (maximum arm SCTT) | cm |
| Fractional lean arm volume (FLAVol) | cm3 |
| Fractional arm volume (AVol) | cm3 |
| Thigha | |
| Femur diaphysis length (FDL) | cm |
| Mid lean thigh circumference (mid lean ThC) | cm |
| Mid-thigh circumference (ThC) | cm |
| Mid-thigh area (MTA) | cm2 |
| Mid-thigh lean area (MTLA) | cm2 |
| Maximum thigh subcutaneous tissue thickness (MTSCTT) | cm |
| Fractional lean thigh volume (FLVol) | cm3 |
| Fractional thigh volume (TVol) | cm3 |
| 2D Abdomen | |
| Abdominal area (AA) | mm2 |
| Standardized abdominal subcutaneous tissue thickness (standardized ASCTT) | mm |
| Maximum abdominal subcutaneous tissue thickness (maximum ASCTT) | mm |
| Cerebellum | |
| Transcerebellar diameter | cm |
| Cerebellum volume | cm3 |
| Adrenals | |
| Adrenal volume | cm3 |
| Adrenal gland length at base | mm |
| Adrenal gland height | mm |
| Adrenal gland width | mm |
| Kidneys | |
| Right kidney volume | cm3 |
| Left kidney volume | cm3 |
| Pancreas | |
| Pancreas circumference | cm |
| Liver | |
| Liver volume | cm3 |
| Lung | |
| Right lung volume | cm3 |
| Left lung volume | cm3 |
| First trimestera | |
| Gestational sac volume | cm3 |
| Embryo volume | cm3 |
| Placenta volume | cm3 |
a In twins, the volumes collected were first-trimester embryo and gestational sac volumes, and in the second and third trimesters thigh (both twins).
Figure 1.
Fractional limb volume. The volumes are based on 50% of the humeral (A) or femoral (B) diaphysis length. Mid-limb measurement eliminates the need for tracing soft tissue borders near the ends of the bone shaft, where acoustic shadowing is more likely to be encountered. Reprinted with permission from Lee et al. (18).
During the standardization process, the expert reviewers determined that the adrenal glands and pancreas could not be measured reliably in all fetuses. In most instances this inability was caused by the orientation of the fetal spine in relation to the transducer. Fetuses who were scanned with the spine at the 12 o’clock position had acoustic shadowing from either the spine and/or lower ribs that obscured the adrenal glands, upper portions of the kidneys, and the pancreas. Furthermore, tissue contrast of the original images was not necessarily optimized for visualizing specific organs.
Quality criteria and quality assurance
Prior to measurement, images and volumes were assessed for whether the region of interest was entirely captured or partially/not captured. They also were assessed for acceptability of measurement. If unacceptable for measurement, the reasons were categorized related to technical factors such as suboptimal image gain (i.e., amplification of an ultrasound signal which makes the image lighter or darker, depending on the echogenicity), excessive fetal motion, or the presence of acoustic shadowing. Each measured image or volume data set was graded: grade A = excellent; B = acceptable; C = unacceptable (Web Table 1). For quality assurance, 10% of scans were selected for repeat measurement (Web Appendix 4). Intra- and interrater agreement were calculated overall and by prespecified gestational age intervals across pregnancy to evaluate potential changes in agreement with advancing gestation (Web Appendix 5).
Data management and integrity
Data entry errors were identified in real time during initial data entry by form-based logic including valid code lists, valid outer ranges, alerts for missing values, missing forms grid, skip logic, and any custom logic checks as deemed necessary. Additional consistency checks were performed after the data entry had been completed and submitted. Regular reports on the data entry and editing processes were generated. Additional quality control procedures include evaluating for intra- and interrater agreement.
Data analysis
Demographic data were summarized as n (%) or mean (within 1 standard deviation). Characteristics were compared for study participants that had 3D available and were included in the Fetal 3D Study and those that did not have 3D available by cohort using χ2 test or Students t test. All analyses were completed with the use of SAS (version 9.4; SAS Institute, Inc., Cary, North Carolina), with a 2-tailed α level of 0.05 considered statistically significant.
RESULTS
Singletons
Out of 2,802 women with a singleton pregnancy, 2,751 (98.2%) were included in the Fetal 3D Study. Of these, 2,435 (88.5%) had at least one 3D ultrasound volume available (while the other 11.5% had only 2D abdomen), and 2,333 (84.8%) had at least one 3D volume that was measured (Figure 2). Maternal characteristics and pregnancy outcomes were similar for women with available 3D when compared with the overall singleton cohort (Table 3).
Figure 2.
Flowchart for inclusion of study participants originally in the National Institute of Child Health and Human Development (NICHD) Fetal Growth Studies, United States, 2009–2013. Singletons and dichorionic twins were included in the NICHD Fetal 3D Study if they had at least one 3D measure available. For singletons, there were 2 body mass index (weight (kg)/height (m)2) cohorts: 19.0–29.9 and 30.0–45.0.
Table 3.
Maternal Characteristics at Enrollment and Pregnancy Outcomes According to Availability of 3D Ultrasound for Singleton Pregnancies (n = 2,751), National Institute of Child Health and Human Development Fetal 3D Study, United States, 2015–2019
| Maternal Characteristic a | NICHD Fetal Growth Studies, Singletons (n = 2,802) | No Record in Fetal 3D Study (n = 51) | Included in the NICHD Fetal 3D Study (n = 2,751) b | ||||||
|---|---|---|---|---|---|---|---|---|---|
| No. | % | Mean (SD) | No. | % | Mean (SD) | No. | % | Mean (SD) | |
| BMI cohortc | |||||||||
| 19.0–29.9 | 2,334 | 83 | 44 | 86 | 2,290 | 83 | |||
| 30–45 | 468 | 17 | 7 | 14 | 461 | 17 | |||
| Racial/ethnic groupd | |||||||||
| White, non-Hispanic | 751 | 27 | 3 | 6 | 748 | 27 | |||
| Black, non-Hispanic | 781 | 28 | 13 | 25 | 768 | 28 | |||
| Hispanic | 803 | 29 | 21 | 41 | 782 | 28 | |||
| Asian, Pacific Islander | 467 | 17 | 14 | 27 | 453 | 16 | |||
| Native US-bornd | 1,914 | 68 | 24 | 49 | 1,890 | 69 | |||
| Age, years | 28.2 (5.5) | 28.4 (5.8) | 28.2 (5.5) | ||||||
| Self-reported heightd | 162.6 (7.1) | 159.5 (7.4) | 162.7 (7.1) | ||||||
| Self-reported prepregnancy weight, kg | 67.3 (14.9) | 65.4 (15.6) | 67.3 (14.9) | ||||||
| Prepregnancy BMIc | 25.4 (5.1) | 25.6 (5.1) | 25.4 (5.1) | ||||||
| Parity | |||||||||
| 0 | 1,320 | 47 | 21 | 41 | 1,299 | 47 | |||
| 1 | 946 | 34 | 16 | 31 | 930 | 34 | |||
| ≥2 | 536 | 19 | 14 | 27 | 522 | 19 | |||
| Marital status | |||||||||
| Not married | 632 | 23 | 12 | 24 | 620 | 23 | |||
| Married or living with partner | 2,082 | 74 | 37 | 74 | 2,045 | 74 | |||
| Divorced, separated, or widowed | 83 | 3 | 1 | 2 | 82 | 3 | |||
| Educationd | |||||||||
| Less than high school graduation | 326 | 12 | 11 | 22 | 315 | 11 | |||
| High school graduation | 513 | 18 | 12 | 24 | 501 | 18 | |||
| Some college or associate degree | 850 | 30 | 16 | 32 | 834 | 30 | |||
| Bachelor’s degree | 645 | 23 | 6 | 12 | 639 | 23 | |||
| Master’s or advanced degree | 466 | 17 | 5 | 10 | 461 | 17 | |||
| Annual family income,$ | |||||||||
| <30,000 | 707 | 29 | 18 | 50 | 689 | 29 | |||
| 30,000–49,999 | 452 | 19 | 7 | 19 | 445 | 19 | |||
| 50,000–74,999 | 311 | 13 | 3 | 8 | 308 | 13 | |||
| 75,000–99,999 | 305 | 13 | 4 | 11 | 301 | 13 | |||
| ≥100,000 | 640 | 27 | 4 | 11 | 636 | 27 | |||
| Health insuranced | |||||||||
| Private/managed care | 1,459 | 52 | 6 | 12 | 1,453 | 53 | |||
| Medicaid; other | 1,125 | 40 | 3 | 6 | 1,122 | 41 | |||
| Currently paid jobs | |||||||||
| 0 | 981 | 35 | 22 | 44 | 959 | 35 | |||
| 1 | 1,700 | 61 | 28 | 56 | 1,672 | 61 | |||
| ≥2 | 118 | 4 | 0 | 0 | 118 | 4 | |||
Table 3.
Continued
| Maternal Characteristic a | NICHD Fetal Growth Studies, Singletons (n = 2,802) | No Record in Fetal 3D Study (n = 51) | Included in the NICHD Fetal 3D Study (n = 2,751) b | ||||||
|---|---|---|---|---|---|---|---|---|---|
| No. | % | Mean (SD) | No. | % | Mean (SD) | No. | % | Mean (SD) | |
| Has at least 1 job or is a full-time student | |||||||||
| No | 836 | 30 | 20 | 40 | 816 | 30 | |||
| Yes | 1,962 | 70 | 30 | 60 | 1,932 | 70 | |||
| Infant sex | |||||||||
| Male | 1,309 | 51 | 5 | 56 | 1,304 | 51 | |||
| Female | 1,249 | 49 | 4 | 44 | 1,245 | 49 | |||
| Pregnancy outcomed | |||||||||
| Live birth ≥20 weeks | 2,562 | 91 | 9 | 18 | 2,553 | 93 | |||
| Fetal death ≥20 weeks | 19 | 1 | 3 | 6 | 16 | 1 | |||
| Miscarriage <20 weeks | 11 | 0 | 2 | 4 | 9 | 0 | |||
| Unknown outcome | 210 | 7 | 37 | 73 | 173 | 6 | |||
| Gestational diabetes mellitus | |||||||||
| No | 2,695 | 96 | 51 | 100 | 2,644 | 96 | |||
| Yes | 107 | 4 | 0 | 0 | 107 | 4 | |||
| Pregnancy-associated hypertensive disorders | |||||||||
| Preeclampsia | 92 | 4 | 0 | 0 | 92 | 4 | |||
| Gestational hypertension | 72 | 3 | 0 | 0 | 72 | 3 | |||
| Unspecified hypertension | 18 | 1 | 0 | 0 | 18 | 1 | |||
| Normotensive | 2,402 | 93 | 9 | 100 | 2,393 | 93 | |||
| Gestational age at delivery, weeks | 39.2 (1.9) | 39.1 (1.2) | 39.2 (1.9) | ||||||
| Preterm delivery <37 weeks | 159 | 6 | 0 | 0 | 159 | 6 | |||
| Birthweight, g | 3,326.0 (529.3) | 3,217.4 (336.2) | 3,326.4 (529.8) | ||||||
| Small for gestational agee | 221 | 9 | 2 | 22 | 219 | 9 | |||
| Appropriate for gestational agef | 2,097 | 83 | 7 | 78 | 2,090 | 83 | |||
| Large for gestational ageg | 223 | 9 | 0 | 0 | 223 | 9 | |||
Abbreviations: BMI, body mass index; NICHD, National Institute of Child Health and Human Development; SD: standard deviation.
a The following variables were missing (US native-born = 5; self-reported height = 4; self-reported weight = 7; BMI = 8; marital status = 5; education = 2; annual family income = 387; health insurance = 218; currently paid job = 3; current student = 4; infant sex = 244; pregnancy-associated hypertensive disorders = 218; preterm delivery = 245; gestational weight gain = 261; gestational age at delivery = 245; birth weight = 251)
b Participants with a record in the Fetal 3D Study could include unmeasurable volumes or only 2D abdominal measurements. Additional details on availability are presented in Table 4.
c Weight (kg)/height (m)2.
d P values < 0.05 calculated by analysis of variance.
e Defined as <10th percentile of birth weight using the Duryea et al. (68).
f Defined as birth weight of ≥10th and ≤90th percentile using the Duryea et al. (68).
g Defined as >90th percentile of birth weight using the Duryea et al. (68).
Table 4.
Continued
| If Unacceptable for Measurement | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Entirely Captured | Region of Interest Partially or Not Captured | Unacceptable for Measurement | Motion Effect b | Mass Effect c | Shadow Effect d | Gain Effect e | |||||||||
| Measure | No. of Records Available in Fetal 3D Study a | No. | % | No. | % | No. | % | No. | % | No. | % | No. | % | No. | % |
| Left kidney volume (total) | 4,996 | 4,532 | 90.7 | 464 | 9.3 | 2,237 | 49.36 | 990 | 44.3 | 64 | 2.9 | 1,616 | 72.2 | 163 | 7.3 |
| Visit 1 | 1,511 | 1,406 | 93.1 | 105 | 7.0 | 603 | 42.89 | 310 | 51.4 | 41 | 6.8 | 385 | 63.9 | 73 | 12.1 |
| Visit 2 | 1,301 | 1,179 | 90.6 | 122 | 9.4 | 568 | 48.18 | 250 | 44.0 | 6 | 1.1 | 431 | 75.9 | 32 | 5.6 |
| Visit 3 | 1,052 | 942 | 89.5 | 110 | 10.5 | 505 | 53.61 | 204 | 40.4 | 5 | 1.0 | 390 | 77.2 | 29 | 5.7 |
| Visit 4 | 794 | 704 | 88.7 | 90 | 11.3 | 395 | 56.11 | 159 | 40.3 | 7 | 1.8 | 292 | 73.9 | 19 | 4.8 |
| Visit 5 | 338 | 301 | 89.1 | 37 | 11.0 | 166 | 55.15 | 67 | 40.4 | 5 | 3.0 | 118 | 71.1 | 10 | 6.0 |
| Liver volume (total) | 4,838 | 4,619 | 95.5 | 219 | 4.5 | 851 | 18.42 | 383 | 45.0 | 120 | 14.1 | 598 | 70.3 | 221 | 26.0 |
| Visit 1 | 1,480 | 1,426 | 96.4 | 54 | 3.7 | 260 | 18.23 | 140 | 53.9 | 39 | 15.0 | 132 | 50.8 | 103 | 39.6 |
| Visit 2 | 1,259 | 1,219 | 96.8 | 40 | 3.2 | 192 | 15.75 | 84 | 43.8 | 32 | 16.7 | 145 | 75.5 | 52 | 27.1 |
| Visit 3 | 1,013 | 963 | 95.1 | 50 | 4.9 | 176 | 18.28 | 73 | 41.5 | 22 | 12.5 | 147 | 83.5 | 35 | 19.9 |
| Visit 4 | 766 | 719 | 93.9 | 47 | 6.1 | 141 | 19.61 | 54 | 38.3 | 20 | 14.2 | 108 | 76.6 | 22 | 15.6 |
| Visit 5 | 320 | 292 | 91.3 | 28 | 8.8 | 82 | 28.08 | 32 | 39.0 | 7 | 8.5 | 66 | 80.5 | 9 | 11.0 |
| Fractional arm volume (total) | 5,552 | 5,338 | 96.2 | 214 | 3.9 | 136 | 2.55 | 31 | 22.8 | 23 | 16.9 | 42 | 30.9 | 13 | 9.6 |
| Visit 1 | 1730 | 1,664 | 96.2 | 66 | 3.8 | 40 | 2.4 | 13 | 32.5 | 1 | 2.5 | 10 | 25.0 | 6 | 15.0 |
| Visit 2 | 1,565 | 1,509 | 96.4 | 56 | 3.6 | 21 | 1.39 | 3 | 14.3 | 4 | 19.1 | 6 | 28.6 | 2 | 9.5 |
| Visit 3 | 1,158 | 1,116 | 96.4 | 42 | 3.6 | 22 | 1.97 | 5 | 22.7 | 5 | 22.7 | 9 | 40.9 | 2 | 9.1 |
| Visit 4 | 807 | 772 | 95.7 | 35 | 4.3 | 34 | 4.4 | 5 | 14.7 | 11 | 32.4 | 13 | 38.2 | 1 | 3.0 |
| Visit 5 | 292 | 277 | 94.9 | 15 | 5.1 | 19 | 6.86 | 5 | 26.3 | 2 | 10.5 | 4 | 21.1 | 2 | 10.5 |
| Fractional thigh volume (total) | 5,599 | 5,287 | 94.4 | 312 | 5.6 | 192 | 3.63 | 31 | 16.2 | 16 | 8.3 | 37 | 19.3 | 4 | 2.1 |
| Visit 1 | 1742 | 1,665 | 95.6 | 77 | 4.4 | 34 | 2.04 | 18 | 52.9 | 6 | 17.7 | 4 | 11.8 | 2 | 5.9 |
| Visit 2 | 1,579 | 1,521 | 96.3 | 58 | 3.7 | 18 | 1.18 | 2 | 11.1 | 1 | 5.6 | 4 | 22.2 | 1 | 5.6 |
| Visit 3 | 1,199 | 1,139 | 95.0 | 60 | 5.0 | 40 | 3.51 | 2 | 5.0 | 4 | 10.0 | 7 | 17.5 | 0 | 0 |
| Visit 4 | 783 | 702 | 89.7 | 81 | 10.3 | 65 | 9.26 | 2 | 3.1 | 4 | 6.2 | 10 | 15.4 | 0 | 0 |
| Visit 5 | 296 | 260 | 87.8 | 36 | 12.2 | 35 | 13.46 | 7 | 20.0 | 1 | 2.9 | 12 | 34.3 | 1 | 2.9 |
a Total counts include the number of women at each visit with measured, unmeasured, and missing volumes values. These numbers vary from the total counts in Web Table 3, which only includes the measured volumes. Denominator for percents in columns 3–5 is the number in column 2; denominator for percents in columns 6–9 is the number in column 5. Note that region of interest partially or not captured and unacceptable for measurement categories overlap (i.e., are not exclusive categories).
b Motion effect: maternal or fetal movement that causes inability to see image clearly.
c Mass effect: lesion or process that causes compression, distortion, or displacement of the structure in an image.
d Shadow effect: refraction and reflection effects at the boundary between the structure and the surrounding tissues. It is often caused by fetal bones.
e Gain effect: gain is amplification of an ultrasound signal which makes the image lighter or darker, depending on the echogenicity. Suboptimal gain can result in decreased visualization of a structure.
The number of women with singleton pregnancies and at least 1 measurement in the Fetal 3D Study was similar across study visits to the number of women with singleton gestations at each visit who had 2D ultrasound data in the NICHD Fetal Growth Studies, except for visit 0 where there were fewer (Figure 3). Fetal 3D Study availability overall and by visit for singleton pregnancies for individual measurements is presented in Web Tables 2 and 3 with additional details, with reasons a volume was unable to be measured provided in Table 4. The volume types with the highest number of measurements available were for the cerebellum, abdomen (liver and kidney), and thigh, indicating prioritization of these anatomical structures per the protocol. In the first trimester (visit 0), the highest number of volumes available were for the embryo because it was more likely that the embryo volume was entirely captured (96.4%) than the placenta and gestational sac volumes, which were often not completely captured in the stored volume sweep (21.9% and 43.8%, respectively). Most of the first-trimester volume data sets were able to be measured if they were entirely captured with only a small percentage being unacceptable for measurement (Table 4). After the first trimester, the highest number of volumes were captured at visit 1 (mean 19.8 (standard deviation, 2.4) weeks) for all volumes except for cerebellar, where the highest number was at visit 2 (27.1 (standard deviation, 2.1) weeks). The number of volumes entirely captured then decreased with advancing gestation, partly due to missed ultrasound visits (including due to delivery) in the original NICHD Fetal Growth Studies, where availability of 2D ultrasound decreased from 97.5% (2,733 out of n = 2,802) in the first trimester to 41.5% (n = 1,162) at visit 5. However, even among ultrasounds where 2D was available, the pattern persisted where the highest number of measured 3D volumes was at visit 1 (except for cerebellum, which peaked at visit 2) and decreased with advancing gestation (Figure 4). In the abdomen, the liver was able to be measured more often than the kidneys. For example, at visit 1, 45.0% of fetuses had a liver volume measurement compared with 30.8% of right kidney and 24.8% of the left kidney (among available 2D) (Figure 4). While the number of volumes were similar (n = 1,480 liver and n = 1,511 for right and left kidneys at visit 1), the liver was slightly more often entirely captured (96%) than the kidneys (93% for both right and left). Also compared with liver volumes, kidney volumes were more likely to be unacceptable for measurement mostly due to shadow or motion effects. Another finding was that right-sided structures (lungs and kidneys) were measured more often than left-sided structures. For fractional limb volumes, both arm and thigh were similarly available.
Figure 3.
Comparison of 2D and 3D ultrasound data availability according to research visit, National Institute of Child Health and Human Development (NICHD) Fetal 3D Study, United States, 2015–2019. Availability means the number of women with singleton gestations at each visit who had 2D ultrasound data in the NICHD Fetal Growth Studies (blue bars) or at least 1 measurement in the Fetal 3D Study (orange bars). Missing and volumes not measured are excluded.
Table 4.
Details for Volume Measurement by Study Visit for Singletons Pregnancies in the National Institute of Child Health and Human Development Fetal 3D Study, United States, 2015–2019
| If Unacceptable for Measurement | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Entirely Captured | Region of Interest Partially or Not Captured | Unacceptable for Measurement | Motion Effect b | Mass Effect c | Shadow Effect d | Gain Effect e | |||||||||
| Measure | No. of Records Available in Fetal 3D Study a | No. | % | No. | % | No. | % | No. | % | No. | % | No. | % | No. | % |
| Placenta volume | 1,812 | 396 | 21.9 | 1,416 | 78.2 | 17 | 4.29 | 5 | 29.4 | 1 | 5.9 | 12 | 70.6 | 4 | 23.5 |
| Embryo volume | 1,812 | 1,747 | 96.4 | 65 | 3.6 | 7 | 0.4 | 6 | 85.7 | 0 | 0 | 3 | 42.9 | 2 | 28.6 |
| Gestational sac volume | 1,812 | 793 | 43.8 | 1,019 | 56.2 | 11 | 1.39 | 6 | 54.6 | 1 | 9.1 | 9 | 81.8 | 0 | 0 |
| Cerebellar volume (total) | 5,568 | 5,125 | 92.0 | 443 | 8.0 | 1,052 | 20.53 | 266 | 25.3 | 71 | 6.8 | 953 | 90.6 | 92 | 8.8 |
| Visit 1 | 1,546 | 1,476 | 95.5 | 70 | 4.5 | 74 | 5.01 | 33 | 44.6 | 9 | 12.2 | 48 | 64.9 | 14 | 18.9 |
| Visit 2 | 1,580 | 1,498 | 94.8 | 82 | 5.2 | 99 | 6.61 | 25 | 25.3 | 5 | 5.1 | 89 | 89.9 | 12 | 12.1 |
| Visit 3 | 1,331 | 1,233 | 92.6 | 98 | 7.4 | 285 | 23.11 | 63 | 22.1 | 17 | 6.0 | 267 | 93.7 | 18 | 6.3 |
| Visit 4 | 829 | 706 | 85.2 | 123 | 14.8 | 408 | 57.79 | 99 | 24.3 | 26 | 6.4 | 380 | 93.1 | 37 | 9.1 |
| Visit 5 | 282 | 212 | 75.2 | 70 | 24.8 | 186 | 87.74 | 46 | 24.7 | 14 | 7.5 | 169 | 90.9 | 11 | 5.9 |
| Right lung volume (total) | 1,507 | 1,449 | 96.2 | 58 | 3.9 | 484 | 33.4 | 353 | 72.9 | 3 | 0.6 | 307 | 63.4 | 22 | 4.6 |
| Visit 1 | 586 | 571 | 97.4 | 15 | 2.6 | 174 | 30.47 | 134 | 77.0 | 1 | 0.6 | 65 | 37.4 | 11 | 6.3 |
| Visit 2 | 397 | 382 | 96.2 | 15 | 3.8 | 96 | 25.13 | 72 | 75.0 | 2 | 2.1 | 66 | 68.8 | 4 | 4.2 |
| Visit 3 | 252 | 243 | 96.4 | 9 | 3.6 | 77 | 31.69 | 48 | 62.3 | 0 | 0 | 62 | 80.5 | 1 | 1.3 |
| Visit 4 | 197 | 183 | 92.9 | 14 | 7.1 | 96 | 52.46 | 69 | 71.9 | 0 | 0 | 79 | 82.3 | 5 | 5.2 |
| Visit 5 | 75 | 70 | 93.3 | 5 | 6.7 | 41 | 58.57 | 30 | 73.2 | 0 | 0 | 35 | 85.4 | 1 | 2.4 |
| Left lung volume (total) | 1,507 | 1,448 | 96.1 | 59 | 3.9 | 571 | 39.43 | 391 | 68.5 | 3 | 0.5 | 383 | 67.1 | 27 | 4.7 |
| Visit 1 | 586 | 571 | 97.4 | 15 | 2.6 | 190 | 33.27 | 139 | 73.2 | 1 | 0.5 | 77 | 40.5 | 11 | 5.8 |
| Visit 2 | 397 | 383 | 96.5 | 14 | 3.5 | 127 | 33.16 | 91 | 71.7 | 0 | 0 | 94 | 74.0 | 5 | 3.9 |
| Visit 3 | 252 | 242 | 96.0 | 10 | 4.0 | 101 | 41.74 | 60 | 59.4 | 0 | 0 | 82 | 81.2 | 2 | 2.0 |
| Visit 4 | 197 | 183 | 92.9 | 14 | 7.1 | 108 | 59.02 | 72 | 66.7 | 2 | 1.9 | 91 | 84.3 | 8 | 7.4 |
| Visit 5 | 75 | 69 | 92.0 | 6 | 8.0 | 45 | 65.22 | 29 | 64.4 | 0 | 0 | 39 | 86.7 | 1 | 2.2 |
| Right kidney volume (total) | 4,996 | 4,503 | 90.1 | 493 | 9.9 | 1,389 | 30.85 | 617 | 44.4 | 62 | 4.5 | 1,133 | 81.6 | 143 | 10.3 |
| Visit 1 | 1,511 | 1,409 | 93.3 | 102 | 6.8 | 476 | 33.78 | 257 | 54.0 | 41 | 8.6 | 330 | 69.3 | 73 | 15.3 |
| Visit 2 | 1,301 | 1,187 | 91.2 | 114 | 8.8 | 371 | 31.26 | 171 | 46.1 | 8 | 2.2 | 330 | 89.0 | 30 | 8.1 |
| Visit 3 | 1,052 | 923 | 87.7 | 129 | 12.3 | 266 | 28.82 | 90 | 33.8 | 7 | 2.6 | 232 | 87.2 | 21 | 7.9 |
| Visit 4 | 794 | 688 | 86.7 | 106 | 13.4 | 194 | 28.2 | 69 | 35.6 | 4 | 2.1 | 169 | 87.1 | 13 | 6.7 |
| Visit 5 | 338 | 296 | 87.6 | 42 | 12.4 | 82 | 27.7 | 30 | 36.6 | 2 | 2.4 | 72 | 87.8 | 6 | 7.3 |
Figure 4.
Availability of 3D fetal volumes according to research visit, National Institute of Child Health and Human Development (NICHD) Fetal 3D Study, United States, 2015–2019. Percent 3D volume availability calculated as the number of women with singleton gestations at each visit who had a measurement for a particular volume divided by the availability of 2D ultrasound data at that visit.
Regarding ability of measurement, fractional lean arm and thigh volumes were less likely to be measured at visit 1 (mean 19.8 (standard deviation, 2.4) weeks) relative to the overall fractional limb volumes: 84.5% (1,377/1,629) for arm and 88.8% (1,455/1,639) for thigh. With advancing gestational age, lean arm and thigh volumes were measured similarly to the overall fractional limb volumes. Maximum arm and thigh SCTT was also not measured as often as the fractional limb volumes at visit 1 but was measured more often than the volumes at subsequent visits. After visit 1, the ability to obtain the maximum abdominal SCTT (ASCTT) was greater than the standardized ASCTT where the measurement was taken in relation to specific fetal anatomy (Web Table 2).
Intrarater and interrater reliability results are presented in Web Tables 4–135 and Web Figures 1–485. For illustration, the overall intrarater and interrater reliability were high for the fractional arm volume, with correlation of 0.99 and 0.98, respectively (Web Tables 12 and 13). Reliability was also high, although slightly less, for the fractional lean arm volume, with correlation of 0.97 and 0.93, respectively. Reliability for the 2D measures calculated from the volume was also high for all dimensions except for the arm SCTT, where the correlation was 0.85 and 0.64 for intrarater and interrater agreement, respectively. The coefficient of variation (CV%) values for intra- and interrater agreement were generally low for most structures, indicating good agreement between the two reviews, specifically 8.7% and 10.7%, respectively for fractional arm volume. CV% values were slightly higher for fractional lean arm volume, 10.6% and 16.5%, mid-arm lean area, 10.6% and 14.3%, and high for arm SCTT, 16.5% and 25.2%, respectively. Similar observation can be made with the Bland and Altman plots, with the mean difference lines near zero, a majority differences inside the 95% confidence band, and a general flat pattern over the range of average differences (Web Figures 23–38). There was no discernable pattern for reliability and agreement differences over gestational age (Web Tables 55–66; Web Figures 174–265).
Findings of high reliability for thigh measurements were similar to arm measurements, with high correlation for all structures (range, 0.96–1.00) except for maximum thigh SCTT, where it was 0.84 and 0.69 for intra- and interrater agreement, respectively (Web Tables 28 and 29). The CV% values for intra- and interrater agreement were generally low for most structures, including 8.7% and 8.6%, respectively, for the fractional thigh volume, with slightly higher CV% values for the fractional lean thigh volume (11.7% and 10.1%) and highest for maximum thigh SCTT (16.9% and 23.3%). Findings for the Bland and Altman plots for the thigh measurements were similar to those for the arm, and again there was no discernable pattern for reliability and agreement differences over gestational age (Web Figures 67–82; Web Tables 117–128; Web Figures 360–455).
When evaluating reliability between historical measurements of the thigh in the NICHD Fetal Growth Studies and measurements in the Fetal 3D Study, again there was high correlation for all structures (range of 0.91–0.96) except for maximum thigh SCTT, where it was 0.57 (Web Appendix 6; Web Tables 32 and 33). The CV% was low for the majority of measurements, including for fractional thigh volume (7.5%), with higher CV% of 10.6% for fractional lean thigh volume, 11.4% for mid-thigh lean area, and 30.6% for thigh SCTT. Findings for the Bland and Altman plots for the thigh measurements between historical and Fetal 3D Study reviewers were similar to those in the Fetal 3D Study, although with slightly more dispersion (Web Figures 99–104 and 456–485). These findings indicate generally good agreement.
Dichorionic twins
Out of 171 women with a dichorionic twin pregnancy, 165 (96.5%) were included in the Fetal 3D Study; all of these had at least one 3D ultrasound volume available, and 152 (92.1%) had at least one 3D volume that was measured (Figure 2). Maternal characteristics and pregnancy outcomes were similar for women with available 3D when compared with the overall twin cohort (Web Table 134). The Fetal 3D Study availability by visit for dichorionic twin pregnancies is presented in Web Table 135. In general, it was more common for at least 1 twin to be measured rather than both twins. The first-trimester volume availability for twins followed the same pattern as singletons, where it was more likely that the embryo volume was available (n = 111 for at least 1 twin, n = 95 for both twins able to be measured) than the placental and gestational sac volumes. The availability of fractional thigh volume for at least 1 twin was generally similar if not slightly higher than that of singletons. Furthermore, overall intrarater and interrater reliability for the thigh measurements were similar between singletons and twins, with high correlation of 1.00 for intra- and interrater agreement for fractional thigh volume and CV% values of 5.2% and 4.6%, respectively (Web Tables 30 and 31).
DISCUSSION
The Fetal 3D Study provides a wealth of 3D data from a diverse pregnant population that is not currently available. Although 3D ultrasound capability is generally available, rendering organ volumes from banked 3D images required training for reviewers and prespecified standard operating procedures for measurement that were developed for the study based on existing literature. The availability of fetal 3D visceral volumes and body composition depended on multiple factors important for planning future fetal 3D studies, including fetal movement and position for the original volume acquisition, that affected initial volume acquisition and image quality (Figure 5).
Figure 5.
Lessons learned.
The increased availability of prioritized anatomical structures, cerebellum, abdomen, and limb volumes, suggests that time likely was a factor in limiting collection of all the 3D data. Duration of the scan should therefore be considered when planning a study. The finding that the majority of first-trimester embryo, placenta, and gestational sac volumes were able to be measured if they were entirely captured, with only a small percentage being unacceptable for measurement, indicates that adequate volume capture is paramount for assessing these structures. Otherwise, at subsequent gestational ages at least one 3D measurement was able to be obtained when 2D was available although suitability for measurement depended on acceptable visualization which was diminished by shadow, motion, mass, and gain effects. For the abdominal structures, the frequent inability to measure the kidneys due to a shadow effect indicates that volumes should be obtained with the fetal spine at 12 or 6 o’clock. Future work is needed to determine the implications of whether a less rigorous (maximum regardless of location) or standardized ASCTT measurement is related to outcomes such as maternal diabetes.
Notably there was generally good agreement between the 2 reviews for the fetal limb structures, although slightly higher for the fractional lean limb volumes and high for the SCTT measures. Our findings are consistent with prior work indicating that fractional arm volume and fractional thigh volume measurements are reproducible during the second and third trimesters of pregnancy (18, 60). Also consistent with our findings, SCTT is known to have increased error in part because the thickness is not continuous around a cross-section of the limb (12). Findings from the Fetal 3D Study have the potential to establish the utility and clinical standard of practice for 3D imaging. For example, diabetes management in pregnancy may be improved by incorporating fetal body composition into monitoring (61). Currently there is no accepted definition for good control in the management of diabetes during pregnancy. However, neonatal fat has been demonstrated to be a measure of maternal glucose control in diabetic mothers. Therefore, real-time identification of accelerated fat deposition in the fetus may be a signal to initiate or adjust medical management to improve outcomes (62). The potential for fetal 3D to assist in pregnancy management has considerable implications for obstetrical practice and public health given the high prevalence of gestational diseases such as preeclampsia (2%–8%) and gestational diabetes (7%) in the United States (63, 64). Given the high prevalence of maternal obesity and excess gestational weight gain in the US population, it is imperative that the effects of these conditions on longitudinal composition of fetal growth be carefully studied.
It should be noted, however, that the time and associated costs might limit current use of 3D in clinical practice. In one small study (n = 50), experts estimated 2–3 minutes per volume for manual fractional limb volume measurement (65). Even under ideal circumstances, an additional 2–3 minutes per patient over the course of a day is cost prohibitive for implementation in a clinical setting. Yet technology is advancing with machine learning, so that automatic calculation of some fetal body composition measurements may become more practical in a clinical environment (66). A semiautomatic tool for calculating fractional limb volume is commercially available, 5 times faster than manual measurement, and may improve precision of EFW formulas to predict birth weight (65, 67). Fetal visceral size and fat may also be important predictors of birth weight, neonatal anthropometrics, and morbidity that may not be otherwise apparent until birth. Combined, these data will respond to critical data gaps in understanding the longitudinal changes in fetal subcutaneous fat, lean body mass, and organ volume in association with gravid diseases.
Supplementary Material
ACKNOWLEDGMENTS
Author affiliations: Epidemiology Branch, Division of Population Health Research, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, United States (Katherine Grantz, Dian He); Division of Women’s and Fetal Imaging, Department of Obstetrics and Gynecology, Baylor College of Medicine, Houston, Texas, United States (Wesley Lee, Lauren Mack, Magdalena Sanz Cortes); Biostatistics and Bioinformatics Branch, Division of Population Health Research, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, United States (Zhen Chen); Department of Obstetrics and Gynecology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States (Stefanie Hinkle); Department of Radiology, Phoenix Children’s Hospital, Phoenix, Arizona, United States (Luis F. Goncalves); Department of Child Health, University of Arizona College of Medicine, Phoenix, Arizona, United States (Luis F. Goncalves); Department of Radiology, University of Arizona College of Medicine, Phoenix, Arizona, United States (Luis F. Goncalves); Department of Radiology, Mayo Clinic, Phoenix, Arizona, United States (Luis F. Goncalves); Department of Radiology, Creighton University, Phoenix, Arizona, United States (Luis F. Goncalves); Department of Obstetrics, Gynecology and Reproductive Sciences, Division of Maternal Fetal Medicine, McGovern Medical School at the University of Texas Health Science Center Houston (UTHealth), Houston, Texas, United States (Jimmy Espinoza); UT Physicians The Fetal Center, Affiliated with Children’s Memorial Hermann Hospital, Houston, Texas, United States (Jimmy Espinoza); The Emmes Company, Rockville, Maryland, United States (Robert E. Gore-Langton, Seth Sherman); The Prospective Group, Fairfax, Virginia, United States (Dian He); Global Center for Asian Women’s Health, and Bia-Echo Asia Centre for Reproductive Longevity and Equality, Yong Loo Lin School of Medicine, National University of Singapore, Singapore (Cuilin Zhang); Department of Obstetrics and Gynecology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore (Cuilin Zhang); and Division of Population Health Research, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, United States (Jagteshwar Grewal).
This research was supported, in part, by the Division of Population Health Research, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), National Institutes of Health; in part, with Federal funds for the Fetal 3D Study (contract numbers: HHSN275201300026I; HHSN275201500002C); and, in part, by the NICHD Fetal Growth Studies—Singletons (contract numbers: HHSN275200800013C; HHSN275200800002I; HHSN27500006; HHSN275200800003IC; HHSN275200800014C; HHSN275200800012C; HHSN275200800028C; HHSN275201000009C). K.L.G., Z.C., and J.G. contributed to this work as part of their official duties as employees of the US Federal Government.
Data generated by this project will be available through the NICHD Data and Specimen Hub (DASH) https://dash.nichd.nih.gov, an NIH-funded/-approved public repository.
The authors thank Nasrin Benion RDMS, Melinda Cheatham RDMS, Emily Crisanti RDMS, Michelle Goepfert RDMS, Maria Hernandez RDMS, Chris Hoang RDMS, Mary Jones RDMS RDCS, Deborah Reid RDMS, Melissa Powell RDMS, and Dr. Mayel Yepez for their efforts in fetal biometric measurements; and the rest of the research team at Baylor College of Medicine, Houston, Texas, as well as the research teams at all participating institutions in the NICHD Fetal Growth Studies who collected the original ultrasound data including, in alphabetical order: Christiana Care Health Systems, Newark, Delaware; Columbia University Medical Center, New York, New York; Fountain Valley Regional Medical Center, Fountain Valley, California; Long Beach Memorial Medical Center, Long Beach, California; Medical University of South Carolina, Charleston, South Carolina; New York Hospital Queens, Flushing, New York; Northwestern University Feinburg School of Medicine, Chicago, Illinois; Saint Peters University Hospital, New Brunswick, New Jersey; The Emmes Company, Rockville, Maryland (data coordinating center); Tufts University, Boston, Massachusetts; University of Alabama, Birmingham, Alabama; University of California, Irvine, Medical Center, Orange, California; and Women and Infants Hospital of Rhode Island, Providence, Rhode Island. The authors thank Drs. Mary D’Alton, Karin Fuchs, and Lawrence Platt for their earlier efforts in helping to develop the ultrasound protocol and perform historical measurements in the NICHD Fetal Growth Studies. The authors also thank The Emmes Company, Rockville, Maryland, for providing data and imaging support for both studies and C-TASC for data support for the NICHD Fetal Growth Studies.
Presented at The Annual Pregnancy Meeting, Society for Maternal-Fetal Medicine, February 6–11, 2023, San Francisco, California.
W.L. has received limited research support from GE HealthCare. L.M. started a new position as Research Program Integrator, Women’s Health Ultrasound, GE HealthCare, after study completion. Measurements for this study were taken using GE software. The other authors report no conflicts.
REFERENCES
- 1. Barker DJ. The fetal and infant origins of adult disease. Editorial. BMJ. 1990;301(6761):1111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Gluckman PD, Hanson MA, Cooper C, et al. Effect of in utero and early-life conditions on adult health and disease. Research support, N.I.H., extramural research support, non-U.S. Gov't review. N Engl J Med. 2008;359(1):61–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. American College of Obstetricians and Gynecologists . ACOG Practice Bulletin, No. 227: Fetal Growth Restriction. Obstet Gynecol. 2021;137(2):e16–e28. [DOI] [PubMed] [Google Scholar]
- 4. American College of Obstetricians and Gynecologists . ACOG Practice Bulletin, No. 216: Macrosomia. Obstet Gynecol. 2020;135(1):e18–e35. [DOI] [PubMed] [Google Scholar]
- 5. Scioscia M, Vimercati A, Ceci O, et al. Estimation of birth weight by two-dimensional ultrasonography: a critical appraisal of its accuracy. Obstet Gynecol. 2008;111(1):57–65. [DOI] [PubMed] [Google Scholar]
- 6. Barel O, Vaknin Z, Tovbin J, et al. Assessment of the accuracy of multiple sonographic fetal weight estimation formulas: a 10-year experience from a single center. Comparative study evaluation studies. J Ultrasound Med. 2013;32(5):815–823. [DOI] [PubMed] [Google Scholar]
- 7. Coomarasamy A, Connock M, Thornton J, et al. Accuracy of ultrasound biometry in the prediction of macrosomia: a systematic quantitative review. BJOG. 2005;112(11):1461–1466. [DOI] [PubMed] [Google Scholar]
- 8. Chauhan SP, Beydoun H, Chang E, et al. Prenatal detection of fetal growth restriction in newborns classified as small for gestational age: correlates and risk of neonatal morbidity. Am J Perinatol. 2014;31(3):187–194. [DOI] [PubMed] [Google Scholar]
- 9. Combs CA, Rosenn B, Miodovnik M, et al. Sonographic EFW and macrosomia: is there an optimum formula to predict diabetic fetal macrosomia? J Matern Fetal Med. 2000;9(1):55–61. [DOI] [PubMed] [Google Scholar]
- 10. Zhang J, Merialdi M, Platt LD, et al. Defining normal and abnormal fetal growth: promises and challenges. Research support, N.I.H., intramural review. Am J Obstet Gynecol. 2010;202(6):522–528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Catalano PM, Tyzbir ED, Allen SR, et al. Evaluation of fetal growth by estimation of neonatal body composition. Research support, U.S. Gov't, P.H.S. Obstet Gynecol. 1992;79(1):46–50. [PubMed] [Google Scholar]
- 12. Ikenoue S, Kasuga Y, Endo T, et al. Newer insights into fetal growth and body composition. Front Endocrinol (Lausanne). 2021;12:708767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Scioscia M, Scioscia F, Vimercati A, et al. Estimation of fetal weight by measurement of fetal thigh soft-tissue thickness in the late third trimester. Ultrasound Obstet Gynecol. 2008;31(3):314–320. [DOI] [PubMed] [Google Scholar]
- 14. Abuelghar W, Khairy A, El Bishry G, et al. Fetal mid-thigh soft-tissue thickness: a novel method for fetal weight estimation. Arch Gynecol Obstet. 2014;290(6):1101–1108. [DOI] [PubMed] [Google Scholar]
- 15. Abramowicz JS, Robischon K, Cox C. Incorporating sonographic cheek-to-cheek diameter, biparietal diameter and abdominal circumference improves weight estimation in the macrosomic fetus. Ultrasound Obstet Gynecol. 1997;9(6):409–413. [DOI] [PubMed] [Google Scholar]
- 16. Kosus N, Kosus A. Can fetal abdominal visceral adipose tissue and subcutaneous fat thickness be used for correct estimation of fetal weight? A preliminary study. J Obstet Gynaecol. 2019;39(5):594–600. [DOI] [PubMed] [Google Scholar]
- 17. Lee W, Mack LM, Sangi-Haghpeykar H, et al. Fetal weight estimation using automated fractional limb volume with 2-dimensional size parameters: a multicenter study. J Ultrasound Med. 2020;39(7):1317–1324. [DOI] [PubMed] [Google Scholar]
- 18. Lee W, Balasubramaniam M, Deter RL, et al. Fractional limb volume—a soft tissue parameter of fetal body composition: validation, technical considerations and normal ranges during pregnancy. Ultrasound Obstet Gynecol. 2009;33(4):427–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Akiba Y, Ikenoue S, Endo T, et al. Differences in fetal fractional limb volume changes in normal and gestational diabetic pregnancies: an exploratory observational study. BJOG. 2021;128(2):329–335. [DOI] [PubMed] [Google Scholar]
- 20. Lee W, Deter R, Sangi-Haghpeykar H, et al. Prospective validation of fetal weight estimation using fractional limb volume. Ultrasound Obstet Gynecol. 2013;41(2):198–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Simcox LE, Myers JE, Cole TJ, et al. Fractional fetal thigh volume in the prediction of normal and abnormal fetal growth during the third trimester of pregnancy. Am J Obstet Gynecol. 2017;217(4):453.e1–453.e12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Steller JG, Gumina D, Driver C, et al. 3D fractional limb volume identifies reduced subcutaneous and lean mass in fetal growth restriction. J Ultrasound Med. 2022;41(7):1623–1632. [DOI] [PubMed] [Google Scholar]
- 23. Catalano PM, Farrell K, Thomas A, et al. Perinatal risk factors for childhood obesity and metabolic dysregulation. Am J Clin Nutr. 2009;90(5):1303–1313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Gluckman PD, Hanson MA, Buklijas T. A conceptual framework for the developmental origins of health and disease. J Dev Orig Health Dis. 2010;1(1):6–18. [DOI] [PubMed] [Google Scholar]
- 25. Larciprete G, Valensise H, Di Pierro G, et al. Intrauterine growth restriction and fetal body composition. Ultrasound Obstet Gynecol. 2005;26(3):258–262. [DOI] [PubMed] [Google Scholar]
- 26. Padoan A, Rigano S, Ferrazzi E, et al. Differences in fat and lean mass proportions in normal and growth-restricted fetuses. Research support, non-U.S. Gov't research support, U.S. Gov't, P.H.S. Am J Obstet Gynecol. 2004;191(4):1459–1464. [DOI] [PubMed] [Google Scholar]
- 27. Kehl RJ, Krew MA, Thomas A, et al. Fetal growth and body composition in infants of women with diabetes mellitus during pregnancy. Comparative study research support, U.S. Gov't, P.H.S. J Matern Fetal Med. 1996;5(5):273–280. [DOI] [PubMed] [Google Scholar]
- 28. Parretti E, Carignani L, Cioni R, et al. Sonographic evaluation of fetal growth and body composition in women with different degrees of normal glucose metabolism. Diabetes Care. 2003;26(10):2741–2748. [DOI] [PubMed] [Google Scholar]
- 29. Larciprete G, Valensise H, Vasapollo B, et al. Fetal subcutaneous tissue thickness (SCTT) in healthy and gestational diabetic pregnancies. Ultrasound Obstet Gynecol. 2003;22(6):591–597. [DOI] [PubMed] [Google Scholar]
- 30. de Santis MS, Taricco E, Radaelli T, et al. Growth of fetal lean mass and fetal fat mass in gestational diabetes. Research support, non-U.S. Gov't. Ultrasound Obstet Gynecol. 2010;36(3):328–337. [DOI] [PubMed] [Google Scholar]
- 31. Konje JC, Okaro CI, Bell SC, et al. A cross-sectional study of changes in fetal renal size with gestation in appropriate- and small-for-gestational-age fetuses. Ultrasound Obstet Gynecol. 1997;10(1):22–26. [DOI] [PubMed] [Google Scholar]
- 32. Kuno A, Hayashi Y, Akiyama M, et al. Three-dimensional sonographic measurement of liver volume in the small-for-gestational-age fetus. J Ultrasound Med. 2002;21(4):361–366. [DOI] [PubMed] [Google Scholar]
- 33. Latini G, De Mitri B, Del Vecchio A, et al. Foetal growth of kidneys, liver and spleen in intrauterine growth restriction: "programming" causing "metabolic syndrome" in adult age. Acta Paediatr. 2004;93(12):1635–1639. [DOI] [PubMed] [Google Scholar]
- 34. Boito SM, Struijk PC, Ursem NT, et al. Assessment of fetal liver volume and umbilical venous volume flow in pregnancies complicated by insulin-dependent diabetes mellitus. BJOG. 2003;110(11):1007–1013. [PubMed] [Google Scholar]
- 35. Britto IS, de Silva Bussamra LC, Araujo Junior E, et al. Fetal lung volume: comparison by 2D- and 3D-sonography in normal fetuses. Arch Gynecol Obstet. 2009;280(3):363–368. [DOI] [PubMed] [Google Scholar]
- 36. Werneck Britto IS, de Silva Bussamra LC, Araujo Junior E, et al. Reference range of fetal lung volume by 3D-ultrasonography using the rotational method (VOCAL). J Perinat Med. 2009;37(2):161–167. [DOI] [PubMed] [Google Scholar]
- 37. Gerards FA, Engels MA, Twisk JW, et al. Normal fetal lung volume measured with three-dimensional ultrasound. Ultrasound Obstet Gynecol. 2006;27(2):134–144. [DOI] [PubMed] [Google Scholar]
- 38. Peralta CF, Cavoretto P, Csapo B, et al. Lung and heart volumes by three-dimensional ultrasound in normal fetuses at 12-32 weeks' gestation. Ultrasound Obstet Gynecol. 2006;27(2):128–133. [DOI] [PubMed] [Google Scholar]
- 39. Gerards FA, Twisk JW, Bakker M, et al. Fetal lung volume: three-dimensional ultrasonography compared with magnetic resonance imaging. Ultrasound Obstet Gynecol. 2007;29(5):533–536. [DOI] [PubMed] [Google Scholar]
- 40. Laudy JA, Janssen MM, Struyk PC, et al. Fetal liver volume measurement by three-dimensional ultrasonography: a preliminary study. Ultrasound Obstet Gynecol. 1998;12(2):93–96. [DOI] [PubMed] [Google Scholar]
- 41. Chang FM, Liang RI, Ko HC, et al. Three-dimensional ultrasound-assessed fetal thigh volumetry in predicting birth weight. Clinical trial comparative study research support, non-U.S. Gov't. Obstet Gynecol. 1997;90(3):331–339. [DOI] [PubMed] [Google Scholar]
- 42. Tedesco GD, Bussamra LC, Araujo Junior E, et al. Reference range of fetal renal volume by three-dimensional ultrasonography using the VOCAL method. Fetal Diagn Ther. 2009;25(4):385–391. [DOI] [PubMed] [Google Scholar]
- 43. Geelhoed JJ, Taal HR, Steegers EA, et al. Kidney growth curves in healthy children from the third trimester of pregnancy until the age of two years. The Generation R Study. Pediatr Nephrol. 2010;25(2):289–298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Buck Louis GM, Grewal J, Albert PS, et al. Racial/ethnic standards for fetal growth: the NICHD fetal growth studies. Am J Obstet Gynecol. 2015;213(4):449.e1–449.e41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Mikolajczyk RT, Zhang J, Betran AP, et al. A global reference for fetal-weight and birthweight percentiles. Lancet. 2011;377(9780):1855–1861. [DOI] [PubMed] [Google Scholar]
- 46. Kiserud T, Piaggio G, Carroli G, et al. The World Health Organization fetal growth charts: a multinational longitudinal study of ultrasound biometric measurements and estimated fetal weight. PLoS Med. 2017;14(1):e1002220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Albert PS, Grantz KL. Fetal growth and ethnic variation. Comment letter. Lancet Diabetes Endocrinol. 2014;2(10):773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Ikenoue S, Akiba Y, Endo T, et al. Defining the normal growth curve of fetal fractional limb volume in a Japanese population. J Clin Med. 2021;10(3):485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Vyas DA, Eisenstein LG, Jones DS. Hidden in plain sight - reconsidering the use of race correction in clinical algorithms. N Engl J Med. 2020;383(9):874–882. [DOI] [PubMed] [Google Scholar]
- 50. Norris KC, Eneanya ND, Boulware LE. Removal of race from estimates of kidney function: first, do no harm. JAMA. 2021;325(2):135–137. [DOI] [PubMed] [Google Scholar]
- 51. Grantz KL, Grewal J, Kim S, et al. Unified standard for fetal growth: the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies. Am J Obstet Gynecol. 2022;226(4):576–587 e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Grantz KL, Grewal J, Kim S, et al. Unified standard for fetal growth velocity: the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies. Am J Obstet Gynecol. 2022;227(6):916–922 e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Papageorghiou AT, Ohuma EO, Altman DG, et al. International standards for fetal growth based on serial ultrasound measurements: the fetal growth longitudinal study of the INTERGROWTH-21st Project. Multicenter study research support, non-U.S. Gov't. Lancet. 2014;384(9946):869–879. [DOI] [PubMed] [Google Scholar]
- 54. Stirnemann J, Villar J, Salomon LJ, et al. International estimated fetal weight standards of the INTERGROWTH-21st Project. Ultrasound Obstet Gynecol. 2017;49(4):478–486. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Buck Louis GM, Grewal J. Clarification of estimating fetal weight between 10–14 weeks gestation, NICHD fetal growth studies. Letter. Am J Obstet Gynecol. 2017;217(1):96–101. [DOI] [PubMed] [Google Scholar]
- 56. Grewal J, Grantz KL, Zhang C, et al. Cohort profile: NICHD Fetal Growth Studies—Singletons and Twins. Int J Epidemiol. 2018;47(1):25–25l. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57. Grantz KL, Grewal J, Albert PS, et al. Dichorionic twin trajectories: the NICHD Fetal Growth Studies. Am J Obstet Gynecol. 2016;215(2):221:e1–221 e16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Zhang C, Hediger ML, Albert PS, et al. Association of maternal obesity with longitudinal ultrasonographic measures of fetal growth: findings from the NICHD Fetal Growth Studies—Singletons. JAMA Pediatr. 2018;172(1):24–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Hediger ML, Fuchs KM, Grantz KL, et al. Ultrasound quality assurance for singletons in the National Institute of Child Health and Human Development Fetal Growth Studies. J Ultrasound Med. 2016;35(8):1725–1733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60. Simcox LE, Higgins LE, Myers JE, et al. Intraexaminer and interexaminer variability in 3D fetal volume measurements during the second and third trimesters of pregnancy. J Ultrasound Med. 2017;36(7):1415–1429. [DOI] [PubMed] [Google Scholar]
- 61. Whitelaw A. Subcutaneous fat in newborn infants of diabetic mothers: an indication of quality of diabetic control. Lancet. 1977;1(8001):15–18. [DOI] [PubMed] [Google Scholar]
- 62. American College of Obstetricians and Gynecologists . ACOG Practice Bulletin, No. 190: Gestational Diabetes Mellitus. Obstet Gynecol. 2018;131(2):e49–e64. [DOI] [PubMed] [Google Scholar]
- 63. American College of Obstetricians and Gynecologists . ACOG Practice Bulletin, No. 222: Gestational Hypertension and Preeclampsia. Obstet Gynecol. 2020;135(6):e237–e260. [DOI] [PubMed] [Google Scholar]
- 64. Martin JA, Hamilton BE, Osterman MJK, et al. Births: final data for 2018. Natl Vital Stat Rep. 2019;68(13):1–47. [PubMed] [Google Scholar]
- 65. Mack LM, Kim SY, Lee S, et al. A novel semiautomated fractional limb volume tool for rapid and reproducible fetal soft tissue assessment. J Ultrasound Med. 2016;35(7):1573–1578. [DOI] [PubMed] [Google Scholar]
- 66. Liu SF, Wang Y, Yang X, et al. Deep learning in medical ultrasound analysis: a review. Engineering-Prc. 2019;5(2):261–275. [Google Scholar]
- 67. Mack LM, Kim SY, Lee S, et al. Automated fractional limb volume measurements improve the precision of birth weight predictions in late third-trimester fetuses. J Ultrasound Med. 2017;36(8):1649–1655. [DOI] [PubMed] [Google Scholar]
- 68. Duryea EL, Hawkins JS, McIntire DD, et al. A revised birth weight reference for the United States. Obstet Gynecol. 2014;124(1):16–22. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.