Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 18.
Published in final edited form as: Obesity (Silver Spring). 2021 Oct;29(10):1676–1683. doi: 10.1002/oby.23237

Practical application of in vivo MRI-based brown adipose tissue measurements in infants

Emily W Flanagan 1, Abby D Altazan 1, Owen T Carmichael 1, Houchun H Hu 2, Leanne M Redman 1
PMCID: PMC9115839  NIHMSID: NIHMS1802547  PMID: 34553508

Abstract

Objective:

The role of brown adipose tissue (BAT) in infant metabolism remains poorly understood, primarily because of the inherent limitation of positron emission tomography/computed tomography imaging to measure BAT, which is not suitable for infants. The aims of this method development study were to assess the feasibility, intra-rater reliability, interscan repeatability, and physiological relevance of measuring BAT in infants using magnetic resonance imaging (MRI).

Methods:

A total of 10 nonsedated infants (mean age, 22.6 [1.3] days old) completed two 3-T MRI exams using chemical-shift-encoded water-fat scans 6.2 (2.8) days apart. Candidate BAT voxels in the supraclavicular region were identified based on fat signal fraction (FSF). The volumes of BAT depots were manually traced, and FSF was calculated. Whole-body fat mass was determined using dual-energy x-ray absorptiometry.

Results:

Images were successfully obtained from 19 of 20 (95%) attempted scans. The mean BAT volume was 5.41 (SD 1.1) cm3, and the mean FSF was 16.41% (SD 3.3%). Intra-rater analysis showed good reliability with no systemic bias (proportional bias for volume: p = 0.19; FSF: p = 0.30). Test-retest for interscan repeatability was good (intraclass correlation coefficients for volume: 0.92, p = 0.001 and intraclass correlation coefficients for FSF: 0.93, p < 0.001). FSF was inversely related to fat-free mass (r = −0.69, p = 0.03).

Conclusions:

This method development study supports the use of MRI to obtain reliable and quantitative measurements of BAT volume in infants.

INTRODUCTION

Developmental origins of obesity begin in utero with the growth and differentiation of adipose tissue. Fetal deposition of fat mass is low during the beginning stages of development and it progresses from around 25 weeks’ gestation until birth (1). The primary role of fetal adipose tissue growth in utero is to protect against postnatal challenges such as low temperature and nutrient availability. Adiposity at birth is highly variable, with percentage of body fat ranging from 7% to 23% (2). Extensive research has suggested that adipose tissue mass at birth is strongly influenced by the maternal metabolic milieu and that it is a representation of future metabolic health (2,3).

Adipose tissue is not uniform in structure, and, therefore, it differs in function. White adipose tissue (WAT) serves as an energy reservoir for lipids and contains a single, large lipid droplet. WAT is the most abundant type of adipose tissue in humans, with depots found primarily in subcutaneous regions and in smaller quantities within the abdominal cavity in and around organs. In infants, the majority of WAT is subcutaneous, with minimal adipose tissue found in the intra-abdominal depots (4). Brown adipose tissue (BAT), on the other hand, is distinct from WAT. The lipid droplets are smaller in size, multilocular, and rich with mitochondria, and they contain a higher capillary density and greater sympathetic innervation (5). Therefore, BAT is a major thermogenic organ with high rates of energy expenditure. In humans, BAT is predominantly located in the supraclavicular, axillary, paraspinous, and perirenal regions. BAT has been found to range from 1 to 170 g in adults (6), but BAT characteristics are richest during infancy and subsequently deteriorate with age (7). Despite morphological changes, most humans preserve some BAT function in adulthood, during which the presence of BAT is associated with a leaner phenotype. However, the physiological relevance of human BAT originates in infancy, with the primary role of supporting thermoregulation after birth. Furthermore, human studies have not yet been conducted to understand how BAT contributes to energy expenditure and metabolism from birth or how it changes throughout childhood.

Research in human infants has been limited by lack of available methodologies. For example, the gold-standard measurement of BAT includes the use of a radioactive isotope tracer (18F-FDG) together with positron emission tomography (PET) and computed tomography (CT), which is not suitable for infants (8). Magnetic resonance imaging (MRI) methods have emerged as a suitable noninvasive imaging technique to identify BAT (9-11). The fat signal fraction (FSF), derived from MRI, is a product of the differential signal from lipid and water. Given the morphological differences from WAT, BAT has been shown to yield a lower FSF (10,12). FSF was effectively compared with PET/CT for the measurement of BAT in adults and children (10).

These imaging techniques for BAT identification have only been available in recent years and have not been well explored in infants. To this end, we conducted a method development study with four aims: 1) to identify the feasibility (data completeness) of using a conventional 3-T MRI to measure BAT in nonsedated infants; 2) to determine intra-rater reliability for manual tracing of BAT depots; 3) to determine test-retest for interscan repeatability across different test days; and 4) to explore the relationship of BAT to fat and fat-free mass in human infants.

METHODS

A total of 10 healthy infants completed two study visits at Pennington Biomedical Research Center in Baton Rouge, Louisiana. The two visits were separated by at least 24 hours but were no more than 10 days apart. Visits were conducted throughout the day, and infants arrived in the postprandial state. The study assessments included anthropometric measurements, body composition via dual-energy x-ray absorptiometry (DXA), and MRI imaging for BAT. The attending parent completed a self-reported questionnaire to assess general health. The first visit was, on average, 118 minutes in length, including the anthropometric assessments, DXA scan, MRI scan, and questionnaire completion. At the second visit, which averaged 57 minutes, the anthropometric measurements and MRI scan were repeated. Sedation was never used to promote sleep during testing. Instead, visits were scheduled around the habitual nap time of each infant. The visit also provided time for the infant to be fed, soothed, and swaddled to sleep. All data were collected between July 2018 and December 2019. The protocol was approved by the Pennington Biomedical Research Center Institutional Review Board. The study was determined to be not more than minimal risk; therefore, written informed consent was obtained from one legal guardian prior to the initiation of testing.

Participants and recruitment

Parents of potential study participants were recruited from targeted advertisements on the Pennington Biomedical Research Center clinical trials LISTSERV email list or Facebook (Facebook Inc., Menlo Park, California), as well as flyers and posters within offices of local midwives, obstetricians, and pediatricians. Infants were eligible if they were no more than 28 days old, born healthy, and were full term. Infants were excluded if they were preterm (<37 weeks), if they had implanted metal or electronic objects that would render MRI procedures unsafe, or if parents were unable to complete two clinic visits within 10 days apart.

Neonatal anthropometrics

Body weight was measured with the infant nude to the nearest 5 g on a calibrated scale (Scale-Tronix Inc., Wheaton, Illinois). Length was measured twice, with the infant’s head in the Frankfort position, using an infantometer (Seca GmbH, Hamburg, Germany) with a stationary headboard and a moveable footboard.

Neonatal body composition

The percentage of body fat mass was assessed using DXA on GE Lunar iDXA (General Electric Healthcare, Chicago, Illinois). Prior to initiating the DXA procedure, the infant was fed and swaddled in a blanket, wearing only a clean diaper, with the arms secured by the side (13-15). The infant was placed on the center of the scanning table while resting and calm. Total scan time was approximately 5 minutes. Fat and fat-free masses were calculated using the percentage of fat mass obtained from the DXA and the measured weight described previously.

Neonatal BAT

MRI scans for BAT were performed using a 3.0-T scanner (Discovery MR750w, General Electric Healthcare). Similar to the DXA procedure, infants were fed, swaddled, and sleeping prior to initiating MRI scanning procedures. Disposable neonatal noise guards (Newmatic Medical, Birmingham, Alabama) specifically designed for infants in MRI environments were gently applied to the ears for noise shielding. A pulse oximeter (Nonin Medical Inc., Plymouth, Minnesota) was applied to the foot, and heart rate and oxygen saturation were continuously monitored during scanning. Research staff were positioned beside the scanner during the procedure to monitor the infant and comfort as needed.

Infants were placed supine in the GE 3.0-T Generalized Encoding Matrix (General Electric Healthcare) head-neck unit coil (neurovascular coil array), which extended over the head and torso of the infant. Padding was used to provide a secure swaddle to prevent movement throughout the scan and to add additional protection from noise. The neurovascular array coil was used to acquire a localizer, IDEAL-IQ, and 3D LAVA scans (General Electric Healthcare). The localizer was used to locate the area of the base of the skull and the dome of the liver. A total of approximately 615 images were acquired on each participant. Acquisition parameters include the following for IDEAL-IQ scans: field of view = 40; matrix = 192 × 192; and section thickness = 2.0 mm with 1.0-mm overlapping sections (section gap = −1.0 mm). Acquisition parameters for the 3D LAVA scans are as follows: field of view = 40; matrix = 192 × 192; and section thickness = 2.0 mm with 1.0-mm overlapping sections (section gap = −1.0 mm). Total scan time with infant positioning was approximately 20 minutes.

Image analysis was performed using Analyze software version 12.0 (AnalyzeDirect, Stilwell, Kansas). Tracing of BAT depots followed a previously described strategy (16). The supraclavicular fat depots, bound by the trapezius muscle posteriorly, the sternocleidomastoid muscle medially, and the clavicle inferiorly, were traced and identified as BAT in the coronal view (Figure 1). When needed, sections in the transverse plane were analyzed to verify BAT depots. On all sections between those boundaries, candidate BAT voxels were identified by thresholding the IDEAL-IQ FSF map in the range between 10% and 50%. Among candidate voxels, finalized BAT voxels were identified by ruling out false positives from other anatomical compartments using 3D LAVA images as a reference. The total volume of all voxels identified as supraclavicular BAT was calculated. One highly experienced image analyzer traced all BAT depots. Intra-rater repeatability was determined from the masked tracing of MRI scans from five distinct infants selected at random.

FIGURE 1.

FIGURE 1

Open in a new tab

IDEAL-IQ image of one infant participant in the (A) coronal view and (B) transverse view for BAT depot tracing. Outlined voxels outlined adjacent to white arrows indicate supraclavicular BAT depots. The images have been modified for visual representation. BAT, brown adipose tissue

Questionnaires

The attending parent completed a general health questionnaire to obtain sociodemographic and anthropometric information of both the biological mother and father, including age, race, weight, information of the infant’s birth (e.g., delivery route, neonatal complications), and infant feeding behaviors.

Statistics

All data were tested for normality and are presented as mean (SD). In order to assess the second aim of reliability of BAT volume and FSF between the two MRI scans and to assess the third aim of intra-rater reliability, intraclass correlation coefficients (ICC) were computed. Intra-rater reliability was evaluated by using Bland-Altman analysis. In order to explore the final aim, physiological relevance of BAT in infants, the relationship between DXA-based measures (percentage of fat and fat-free mass) and MRI-based measures (BAT volume and mean FSF within the BAT) were estimated with Pearson correlation. The average value for BAT volume and mean BAT FSF between the two repeated scans were used for correlation analyses. All analyses were performed using SPSS version 24 (SPSS Inc., Chicago, Illinois) with α ≤ 0.05 set as the predetermined level of significance.

RESULTS

A total of 10 full-term infants between the ages of 14 and 27 days old at the first visit (mean 22.6 [4.1] days) participated in the study. The majority of the infants were male (n = 7), White (n = 9), and all infants were breastfed at the time of the study visits. The average age of the biological mother and father was 32.1 (5.7) and 34.4 (6.3) years, respectively. Participant characteristics of infants are shown in Table 1.

TABLE 1.

Summary of infant characteristics, n = 10

Mean ± SD Range
Anthropometrics
 Weight (kg) 4.41 ± 4.0 3.5-4.8
 Length (cm) 53.7 ± 1.7 50.9-56.1
Body composition
 Fat mass (g) 891.6 ± 114.0 740.9–1,096.2
 Fat mass (%) 20.8 ± 1.8 18.3-24.0
 Fat-free mass (kg) 3.39 ± 0.31 2.73-3.86
BAT
 BAT volume (cm3)* 5.4 ± 1.1 4.3-7.8
 FSF (%)* 16.41 ± 3.3 12.6-22.0
Open in a new tab

Abbreviations: BAT, brown adipose tissue; FSF, fat signal fraction.

*

Mean volume and FSF from 19 scans are presented.

Data completeness

For all scans, the infant was successfully positioned supine in the head coil, and the MRI procedures were successfully executed. Of the expected 20 MRI scans, 19 (95%) were obtained and produced valid images for analysis. During one scan, which was unusable, an MRI machine malfunction led to acquisition of images that were unanalyzable.

Intra-rater reliability

For five distinct sets of images, the volume of BAT measured during the first analysis was 4.9 (1.0) cm3 and was 5.5 (0.8) cm3 during the second analysis. The mean coefficient of variation for BAT volume in the supraclavicular depot was 8.8% between two analyses. The subsequent FSF of BAT measured during each analysis was 17.7% (4.0%) and 17.4% (3.9%). As shown in Figure 2A,B, there was excellent reliability between the two independent analyses for both BAT volume (r = 0.85, p = 0.001) and FSF (r = 0.92, p < 0.001). The Bland-Altman plot shows that the proportional bias was not significant across review for neither BAT volume (p = 0.19) nor FSF (p = 0.30).

FIGURE 2.

FIGURE 2

Open in a new tab

Intra-rater reliability for BAT (A) volume and (B) FSF and proportional bias for (C) volume and (D) FSF, represented by Bland-Altman plots. BAT, brown adipose tissue; FSF, fat signal fraction

Interscan repeatability analysis

A moderate-to-excellent degree of reliability between the two scans was observed for BAT depot volume (Figure 3A), and a good-to-excellent degree of reliability was observed for FSF (Figure 3B). At the first visit, BAT volume was 5.41 (0.88) cm3 and was 5.50 (1.30) cm3 at the second visit. The average measured ICC for BAT volume was 0.92, with a 95% confidence interval from 0.65 to 0.98 (F[8,8] = 12.60, p < 0.001). Average FSF at the first visit was 16.64% (3.31%) and 16.43% (3.41%) at the second visit. The average measured ICC for FSF was 0.93, with a 95% confidence interval from 0.82 to 0.99 (F[162,972] = 5.78, p < 0.001).

FIGURE 3.

FIGURE 3

Open in a new tab

Relationship between (A) BAT volume and (B) BAT FSF obtained during two clinic visits conducted within 10 days apart. BAT, brown adipose tissue; FSF, fat signal fraction

Physiological relevance

Taking into account all 19 scans, the average BAT volume was 5.41 (1.1) cm3 with a FSF of 16.41% (3.3%). FSF was inversely related to fat-free mass (r = −0.69, p = 0.03) and positively related to percentage of fat mass (r = 0.64, p = 0.05; Figure 4A,B). There was a nonsignificant inverse relationship between FSF and body weight (r = −0.56, p = 0.08). After controlling for age of the infant (in days), the relationship between BAT depot FSF and fat-free mass remained strong; however, because of the sample size, it was no longer significant (r = −0.66, p = 0.06).

FIGURE 4.

FIGURE 4

Open in a new tab

Relationship between infant fat signal fraction and (A) fat free mass and (B) fat mass percentage

DISCUSSION

Reliable and noninvasive imaging techniques to quantify BAT are vital to understand the intrauterine development of BAT and how BAT structure and function adapt in the early postnatal period and across the life course. This method development study successfully demonstrated that the MRI IDEAL-IQ, together with 3D LAVA images, can feasibly and reliably measure BAT volume and FSF of BAT in neonates. Until now, to our knowledge, no studies have shown the reliability of measuring BAT in neonates across two distinct scans. We also demonstrate the physiological relevance of FSF to body composition in neonates.

Compared with PET/CT, which relies on tissue uptake of a radioactive tracer (17), FSF relies on the morphological differences between BAT and WAT to produce differential signals detected by MRI (10). The use of FSF in infants to identify BAT has been confirmed by Hu et al. (18) through tissue histology in a postmortem infant and has since been used to determine BAT in nonsedated infants (16,19). A lower FSF in the basal state is a result of less triglyceride content and may additionally indicate greater mitochondrial density and vascularization. During activation (e.g., during a mild cold stimulation), it has been hypothesized that a lower FSF would indicate a reduction in intracellular triglyceride (20). Therefore, using FSF in thermoneutrality, together with cold stimulation, may provide additional insight into BAT functionality. In the present study, BAT was measured in a temperature-controlled room with infants comfortably swaddled, and, therefore, BAT is not expected to have been activated. Nonetheless, key limitations for the use of MRI to quantify infant BAT should be acknowledged. It is important to note that FSF reflects tissue histology, not activity. Therefore, a lower FSF during BAT activation (e.g., cold stimulation) likely reflects lower intracellular triglyceride content. However, FSF does not provide information on the metabolic fate of the fatty acids within the tissue (e.g., oxidation, re-esterification, release into circulation).

The acquisition protocol used in the present study had a 95% success rate for acquiring the desired images. It should be noted that the one scan produced images that were unsuitable for analysis. Importantly, it was not the result of a procedural flaw or noncompliance of the infant with the scanning procedure. The high level of data completeness is likely the result of many procedural considerations deployed throughout the trial. For instance, all scans were completed during the habitual nap time of the infant. This required that visit scheduling and length were flexible in order to allow the infant to be fed and swaddled to sleep. Second, infant comfort was a high priority throughout the scanning procedure. A staff member was present in the imaging room beside the scanner during all scans to ensure that the infant was calmly resting or sleeping. Because the infant was positioned inside the head coil, it was difficult to visually monitor vital signs. A pulse oximeter was applied to the infant’s foot, which allowed for constant observation of heart rate and oxygen saturation. Owing to the fact that heart rate begins to steadily rise upon the early stages of wakening, heart-rate observation during scanning allowed the staff member to detect changes in the infant’s comfort more easily. If the infant woke or showed early signs of discomfort, the staff member would signal to pause the scan and soothe the infant back to sleep without needing to remove the infant from the scanning table. The majority of wakening episodes occurred during the initial localization scans; the abrupt noise from the MRI scanner startled the infant. The use of the neonatal noise guards, together with the support pads tucked around the ears, assisted to dampen the noise. Stopping criteria for suspending the scan included noticeable early stages of wakening or discomfort and crying. Scans were resumed once the infant was soothed back to sleep, as observed visually and by declines in infant heart rate.

There are other important methodological details to consider when attempting to measure BAT in infants. One important consideration is the selection of the BAT depot for measurement. In infants, BAT is predominantly located in four regions of the body: paraspinous, supraclavicular, axillary, and perirenal. The supraclavicular region is most commonly investigated in studies involving children and adults. This is due in part to the ease of identification and that the supraclavicular region is retained across the life course (21,22). In order to study the longitudinal effects of BAT and for comparison with adult populations, it is important to maintain consistency in the chosen BAT depot. Second, because BAT is not automatically identified with MRI, there are several considerations that should be applied to postimaging processing to ensure valid outcomes. The degree of FSF applied to identify BAT regions in the region of interest must be determined a priori. The decision to perform manual tracing of BAT depots versus computer-generated results must be weighed. Manual tracing of BAT voxels is more time consuming and prone to inter-rater differences. However, use of computer software, although streamlined and autonomous, may falsely identify other anatomical structures in addition to BAT that produce a similar FSF (data not shown). Finally, recent evidence has suggested a role for BAT to contribute to the transient rise in energy expenditure following a meal (i.e., diet-induced thermogenesis) (23), which highlights the potential importance of fasted scans or standardizing meal consumption prior to imaging. Neonates (i.e., infants <30 days old) are a vulnerable population in which fasted visits are unethical, and it is progressively more difficult to standardize meals. As previously noted, in the present study, all infants were imaged in the postprandial state. Future studies should consider the influence of BAT on diet-induced thermogenesis and attempt to standardize infant meal timing prior to imaging while understanding that the needs of infants are more complex than adults.

There are several notable limitations in the current study design that should be considered in future research. First, our cohort is limited by a small and relatively homogenous sample of 10 infants. Given the nature of the method development study, we did not enroll equal numbers of male infants and female infants. There is evidence to support a potential sex-difference in BAT during adulthood (6); therefore, future studies may benefit from enrolling both sexes equally. Our approach to identify BAT included manual tracings of individual voxels with a FSF between 10% and 50%, examining images in multiple planes, and comparing traced images against anatomical reference scans, thus minimizing false identification of other tissues as BAT. Previous studies applied a higher FSF parameter (20%-60%), which may have identified adipose tissue that more closely resembles WAT as BAT. Additionally, the use of automated software to position seed bubbles around regions of interest to identify tissues (24) may have resulted in other tissues proximal to the BAT being assumed to be BAT as well (19,25). As a result of our more conservative approach, FSF observed in the supraclavicular depot is less than previously reported in healthy, similar-aged infants (16,19,24,25) but comparable with infants born with hypoxic-ischemic encephalopathy who underwent hypothermia therapy (26). In the present study, we did not obtain a full body MRI scan; therefore, we are unable to compare the total volume of BAT with total adipose tissue volume. Owing to our uncertainty in the ability to successfully acquire MRI scans in healthy, non-sedated infants, we elected to evaluate fat mass using DXA. Finally, MRI has been used to identify alternative physiologically relevant BAT depots such as the intrascapular region. However, given the positioning of the infant (i.e., swaddled, supine) and need for a shorter scan time, the identification of other depots is inherently difficult in this population. Therefore, the tradeoff between specificity and infant comfort (scan time, infant positioning) remains a fundamental limitation.

Young neonates lack adequate musculature for shivering; therefore, heat is produced through thermogenesis of BAT. Increased energy expenditure together with data supporting relationships between BAT activation and a more-lean phenotype in adults (22) has ignited investigation into BAT as a potential therapeutic for weight loss. However, because BAT originates in infancy, it is important to understand the physiological relevance during this life stage. There are no studies, to our knowledge, that have investigated the role of BAT on energy expenditure in infants. Although similar relationships have been observed between BAT and adiposity or BMI in children (16,27), to our knowledge, this is the first study to elucidate the relationship between fat mass and BAT in young infants. To this end, we also observed a positive relationship between fat-free mass and FSF. It will be important to understand whether BAT is influenced by uterine factors, including the maternal metabolic milieu, gestational weight gain, or maternal exercise behaviors. Recent evidence has suggested that BAT may be influenced by thermogenic metabolites within breastmilk (28). Postnatal nutrition should be investigated as a potential stimulator of BAT during neonatal development.

CONCLUSION

In adults, activation of BAT can influence whole-body metabolism and it is associated with superior health phenotypes. A missing link remains between neonatal BAT and the relationship to adult BAT presence. In the present study, we demonstrate that BAT can be measured with high success and excellent reliability. Additionally, we provide support for relationships between intrauterine BAT development with adiposity and fat-free mass. Now that reliable methodologies to measure and analyze BAT in human neonates have been established, research should begin to disentangle the relationships of BAT on energy expenditure and to determine influences from preconception parental health, the intrauterine environment, and postnatal nutrition. We can now begin to elucidate the role of BAT as a protective phenotype against the development of childhood obesity.

Study Importance.

What is already known?

  • Brown adipose tissue (BAT) is a major thermogenic organ and is associated with a more-lean phenotype in adults. However, the physiological relevance of BAT is during infancy to produce heat in the postnatal environment.

  • Positron emission tomography/computed tomography is most often used to measure BAT in adults, and this method is not suitable during infancy. Magnetic resonance imaging (MRI) has recently emerged as a suitable alternative to measure BAT fat signal fraction in infants and children.

What does this study add?

  • MRI can be used with a high success rate and reproducible results when a standard protocol is followed.

  • A relationship is observed between BAT and fat and fat-free mass during infancy, highlighting the physiological relevance during this life stage.

How might these results change the direction of research?

  • Publication of practical applications to guide measurement of BAT in infants with MRI allows for expansion of BAT research during infancy. Further investigation in this field is crucial to understand the role of BAT in the development or prevention of obesity.

Funding information

This research was funded in part by the Pennington/Louisiana Nutrition Obesity Research Center (grant P30 DK072476-10). LMR and EWF are funded in part by grants R01 NR017644 and R01 DK124806.

Footnotes

CONFLICT OF INTEREST

The authors declared no conflict of interest.

REFERENCES

  • 1.Ziegler EE, O'Donnell AM, Nelson SE, Fomon SJ. Body composition of the reference fetus. Growth. 1976;40(4):329–341. [PubMed] [Google Scholar]
  • 2.Toro-Ramos T, Paley C, Pi-Sunyer FX, Gallagher D Body composition during fetal development and infancy through the age of 5 years. Eur J Clin Nutr. 2015;69(12):1279–1289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Ong KK. Size at birth, postnatal growth and risk of obesity. Horm Res. 2006;65(suppl 3):65–69. [DOI] [PubMed] [Google Scholar]
  • 4.Harrington TA, Thomas EL, Frost G, Modi N, Bell JD Distribution of adipose tissue in the newborn. Pediatr Res. 2004;55(3):437–441. [DOI] [PubMed] [Google Scholar]
  • 5.Cedikova M, Kripnerova M, Dvorakova J, et al. Mitochondria in white, brown, and beige adipocytes. Stem Cells Int. 2016;2016:6067349. doi: 10.1155/2016/6067349 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Cypess AM, Lehman S, Williams G, et al. Identification and importance of brown adipose tissue in adult humans. N Engl J Med. 2009;360(15):1509–1517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Heaton JM. The distribution of brown adipose tissue in the human. J Anat. 1972;112(Pt l):35–39. [PMC free article] [PubMed] [Google Scholar]
  • 8.Wu M, Junker D, Branca RT, Karampinos DC Magnetic resonance imaging techniques for brown adipose tissue detection. Front Endocrinol (Lausanne). 2020;11:421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Chen YC, Cypess AM, Chen YC, et al. Measurement of human brown adipose tissue volume and activity using anatomic MR imaging and functional MR imaging. J Nucl Med. 2013;54(9):1584–1587. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hu HH, Perkins TG, Chia JM, Gilsanz V. Characterization of human brown adipose tissue by chemical-shift water-fat MRI. AJR Am J Roentgenol. 2013;200(1):177–183. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Borga M, Virtanen KA, Romu T, et al. Brown adipose tissue in humans: detection and functional analysis using PET (positron emission tomography), MRI (magnetic resonance imaging), and DECT (dual energy computed tomography). Methods Enzymol. 2014;537:141–159. doi: 10.1016/B978-0-12-411619-1.00008-2 [DOI] [PubMed] [Google Scholar]
  • 12.Karampinos DC, Weidlich D, Wu M, Hu HH, Franz D. Techniques and applications of magnetic resonance imaging for studying brown adipose tissue morphometry and function. Handb Exp Pharmacol. 2019;251:299–324. [DOI] [PubMed] [Google Scholar]
  • 13.Fields DA, Demerath EW. Relationship of insulin, glucose, leptin, IL-6 and TNF-alpha in human breast milk with infant growth and body composition. Pediatr Obes. 2012;7(4):304–312. doi: 10.1111/j.2047-6310.2012.00059.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Godang K, Qvigstad E, Voldner N, et al. Assessing body composition in healthy newborn infants: reliability of dual-energy x-ray absorptiometry. J Clin Densitom. 2010;13(2):151–160. [DOI] [PubMed] [Google Scholar]
  • 15.Gallo S, Vanstone CA, Weiler HA. Normative data for bone mass in healthy term infants from birth to 1 year of age. J Osteoporos. 2012;2012:672403. doi: 10.1155/2012/672403 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hu HH, Yin L, Aggabao PC, Perkins TG, Chia JM, Gilsanz V. Comparison of brown and white adipose tissues in infants and children with chemical-shift-encoded water-fat MRI. J Mag Reson Imaging. 2013;38(4):885–896. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Ouellet V, Labbe SM, Blondin DP, et al. Brown adipose tissue oxidative metabolism contributes to energy expenditure during acute cold exposure in humans. J Clin Invest. 2012;122(2):545–552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Hu HH, Tovar JP, Pavlova Z, Smith ML, Gilsanz V. Unequivocal identification of brown adipose tissue in a human infant. J Magn Reson Imaging. 2012;35(4):938–942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Rasmussen JM, Entringer S, Nguyen A, et al. Brown adipose tissue quantification in human neonates using water-fat separated MRI. PLoS One. 2013;8(10):e77907. doi: 10.1371/journal.pone.0077907 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Oreskovich SM, Ong FJ, Ahmed BA, et al. MRI reveals human brown adipose tissue is rapidly activated in response to cold. J Endocri Soc. 2019;3(12):2374–2384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Au-Yong IT, Thorn N, Ganatra R, Perkins AC, Symonds ME. Brown adipose tissue and seasonal variation in humans. Diabetes. 2009;58(11):2583–2587. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, et al. Cold-activated brown adipose tissue in healthy men. New Engl J Med. 2009;360(15):1500–1508. [DOI] [PubMed] [Google Scholar]
  • 23.Hibi M, Oishi S, Matsushita M, et al. Brown adipose tissue is involved in diet-induced thermogenesis and whole-body fat utilization in healthy humans. Int J Obes (Lond). 2016;40(11):1655–1661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ponrartana S, Aggabao PC, Chavez TA, Dharmavaram NL, Gilsanz V. Changes in brown adipose tissue and muscle development during infancy. J Pediatr. 2016;173:116–121. [DOI] [PubMed] [Google Scholar]
  • 25.Entringer S, Rasmussen J, Cooper DM, et al. Association between supraclavicular brown adipose tissue composition at birth and adiposity gain from birth to 6 months of age. Pediatr Res. 2017;82(6):1017–1021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Hu HH, Wu TW, Yin L, et al. MRI detection of brown adipose tissue with low fat content in newborns with hypothermia. Magn Reson Imaging. 2014;32(2):107–117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Drubach LA, Palmer EL 3rd, Connolly LP, Baker A, Zurakowski D, Cypess AM. Pediatric brown adipose tissue: detection, epidemiology, and differences from adults. J Pediatr. 2011;159(6):939–944. [DOI] [PubMed] [Google Scholar]
  • 28.Wolfs D, Lynes MD, Tseng Y-H, et al. Brown fat-activating lipokine 12,13-diHOME in human milk is associated with infant adiposity. J Clin Endocrinol Metab. 2021;106:e943–e956. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES