Current and potential methods to assess kidney structure and morphology in term and preterm neonates

Abstract

After birth, the kidney structure in neonates adapt to the functional demands of extrauterine life. Nephrogenesis is complete in the third trimester, but glomeruli, tubuli, and vasculature mature with the rapidly increasing renal blood flow and glomerular filtration. In preterm infants, nephrogenesis remains incomplete and maturation is slower and may be aberrant. This structural and functional deficit has life‐long consequences: preterm born individuals are at higher risk for chronic kidney disease and arterial hypertension later in life. This review assembles the literature on existing and potential methods to visualize neonatal kidney structure and morphology and explore their potential to longitudinally document the developmental deviation after preterm birth. X‐rays with and without contrast, fluoroscopy and computed tomography (CT) involve relevant ionizing radiation exposure and, apart from CT, do not provide sufficient structural details. Ultrasound has evolved into a safe and noninvasive high‐resolution imaging method which is excellent for longitudinal observations. Doppler ultrasound modes can characterize and quantify blood flow to and through the kidneys. Microvascular flow imaging has opened new possibilities of visualizing previously unseen vascular structures. Recent advances in magnetic resonance imaging display renal structure and function in unprecedented detail, but are offset by the logistical challenges of the imaging procedure and limited experience with the new techniques in neonates. Kidney biopsies visualize structure histologically, but are too invasive and remain anecdotal in newborns. All the explored methods have predominantly been examined in term newborns and require further research on longitudinal structural observation in the kidneys of preterm infants.

1. INTRODUCTION

The structure and function of the kidney in newborns differ significantly from adults and develop rapidly over the first months and years of life. Nephrogenesis, the formation of new nephrons, is complete in utero at around 34–36 weeks of gestation, but the size of the glomeruli in a full‐term newborn is only about 40%, and the length of tubules only 10% that of an adult (Hinchliffe et al., 1991). The glomerular volumes are heterogenous across the cortex because of the centrifugal maturation: glomeruli in the perimedullary cortical area grow and mature earlier than those in the outer subcapsular area where the youngest nephrons form (the so‐called nephrogenic zone) (Minuth, 2021). After birth, renal blood flow and blood pressure increase rapidly while peripheral vascular resistance falls sharpy, which leads to more than doubling of the glomerular filtration rate (GFR) from around 20 mL/kg/1.73 m² in a term newborn in the first weeks of life. GFR then continues to steadily rise until reaching adult levels in the second year of life (Iacobelli & Guignard, 2021). Even though no new nephrons form after birth, the functional demands during the early period of life result in the increase of glomerular volumes, filtration area and tubular length, and lengthening and branching of the vascular tree (Saint‐Faust et al., 2014; Segar, 2017; Sequeira Lopez & Gomez, 2011). In infants born prematurely, much of the organogenesis, maturation, and growth of the kidneys take place ex utero, often during periods of illness associated with prematurity. In autopsy studies, the total number of glomeruli was found to be proportional to birth weight, and the kidneys only have a limited capacity for postnatal nephrogenesis after an early birth (Manalich et al., 2000; Rodriguez et al., 2004). Many of the newly formed glomeruli in preterm infants are morphologically abnormal, with a narrower nephrogenic zone (Black et al., 2013). Animal studies have confirmed the association between preterm birth and low total nephron numbers, but also with reduced glomerular vascularization (Staub et al., 2017). Typical medical complications and treatments during the admission to the neonatal intensive care unit (NICU) such as ventilation, sepsis, acute kidney injury, patent ductus arteriosus (PDA), nephrotoxic drugs, and extrauterine growth restriction, further impair the structural and functional development of the kidneys (Iacobelli & Guignard, 2022). According to the Brenner hypothesis, these early life events with the resulting lifelong nephron deficit lead to compensatory glomerular hyperfiltration which in turn leads to glomerular damage and further nephron loss over time. This sets the preterm born individual up for decreased renal function and higher blood pressure later in life (Brenner et al., 1988). Epidemiological studies show clearly how preterm birth and low birth weight are important risk factors for chronic kidney disease and cardiovascular disease from early adulthood (Risnes et al., 2021; Ueda et al., 2014; White et al., 2009). Despite the increasing knowledge of the renal consequences of preterm birth later in life, there is a distinct lack of methods that could demonstrate in vivo and in a longitudinal fashion how structure (defined as the macro‐ and microscopic parts of an organ) and morphology (the relationship of said structures to each other or to physical properties of an organ) deviate from the normal newborn trajectory after an infant is born preterm. Visualization of morphological structures and associated function in preterm and term infants could advance the understanding of the pathophysiology behind the differences and identify individuals at particular risk of later renal sequelae. Eventually, this knowledge could lead to treatments that are gentler on the kidneys and avoid, or at least reduce the impact of these late consequences of preterm birth.

The aim of this literature review is to review imaging and other methods that allow the in vivo assessment of kidney structure and morphology in preterm and term infants. Some emphasis lies on the technical aspects of each method and how the individual modalities are applicable to the small size, transitioning physiology, and still maturing organ systems in the neonatal population.

2. METHODS

We systematically searched the following databases to identify articles for this literature review: Medline, Embase, Medline, CINAHL, Google Scholar. The searches were conducted in August and September 2022. No language restriction was used; however, the years were capped from 2000 to 2022 to reflect contemporary methods (unless an older article referred to fundamental principles of, or findings for a particular method).

The search terms aimed at identifying relevant articles on imaging or other methods that can display morphological structure of the kidneys in neonates. The MeSH term identified for Medline and Embase were adapted to corresponding terms in CINAHL and used in Google Scholar. Each search term was supplemented with relevant free text terms. The database searches were complemented with manual review of the reference list of relevant articles. Search strategies are specified in Appendix A. Inclusion criteria were as follows: (a) Article related to the use of imaging methods or biopsies to display kidney structure or morphology in newborns (or in other patient population with relevance for newborns), (b) Publication format of original scientific articles, case reports, editorials, literature reviews or book chapters, (c) Full text were available in English, German, French, or Spanish. The author acknowledges that, despite the broad search strategy, relevant articles missing keywords related to kidney images or structure might have been missed.

From an initial 2,651 citations, a total of 311 articles were identified for full text screening. Among those, 76 were deemed relevant to this review and included into the following paragraphs subdivided according to methods of display for kidney structure and morphology: x‐ray with or without contrast, computed tomography (CT), nuclear medicine imaging, ultrasound, magnetic resonance imaging (MRI), and biopsy.

3. RESULTS

3.1. X‐ray with and without contrast

3.1.1. Technical notes

Radiographs or x‐ray images are generated when high‐energy ionizing electromagnetic radiation that travels through the body and are absorbed differentially depending on tissue density. Dense tissue (such as bone) absorbs most of the x‐ray beam, whereas air‐filled structures such as the lungs let x‐rays readily pass. A detector plate on the other side of the patient captures the x‐rays and transforms the information into a gray scale image, with dense structures displayed in brighter, and less dense tissue in darker shades of gray.

3.1.2. Native x‐ray and intravenous pyelograms

Kidney, ureter, and bladder (KUB) radiographs are optimized for the softer organ structures of the reno‐urinary tract, but only give the approximate size and outline of the kidneys and are therefore not commonly used to display kidney structures. KUB radiographs are, however, the baseline images for contrast studies such as intravenous urograms (IVUs, or excretory urograms) (Er A. Abdomen (KUB view), 2020). For these, iodinated contrast agent injected intravenously outlines the kidneys, ureters, and bladder on plain x‐ray film as it absorbs the radiation passing through the field. Some of the earliest imaging studies to measure kidney size were performed with this technique (Hodson et al., 1962). Today, there is no indication for IVUs in the neonatal period, as even in healthy term newborns with normal renal function IVUs may fail to outline the kidneys in the first days of life. Worryingly, intravenous use of iodinated contrast agents can cause hypothyroidism in neonates (Ahmet et al., 2009). Aggravation of renal venous thrombosis or medullary necrosis are rare complications described after IVUs in young infants (de Bruyn & Marks, 2008).

3.1.3. Fluoroscopy

Fluoroscopy uses x‐rays and a fluorescent screen for real time diagnostic images of moving structures or movement of contrast agent (e.g., blood flow after intravenous contrast injection, gastrointestinal passage after contrast ingestion). The radiation exposure for a single image is comparatively lower than for a conventional static x‐ray, but the repeated short bursts of radiation over an entire fluoroscopic imaging procedure sum up to a higher total radiation dose, particularly in complex diagnostic situations (Wambani et al., 2014). Micturating cystourethrograms (MCUG) or voiding cystourethrograms (VCUG) involve the insertion of a catheter and instillation of contrast into the bladder. This imaging technique is traditionally used to visualize and grade vesicoureteral reflux (VUR) where the contrast medium backfills the ureter(s) and pelvi‐calyceal system from the bladder (Figure 1). In male patients, the lateral views are mandatory to evaluate the presence or absence of posterior urethral valves, typically seen as an abrupt change in caliber of the urethra. MCUG remains the most commonly used and gold standard imaging modality for diagnosing VUR, but newer techniques of contrast enhanced or low flow Doppler ultrasound (Gordon & Riccabona, 2003; Yoo et al., 2020) and magnetic resonance urography (MRU; Avni et al., 2001; Rodriguez et al., 2001) now offer alternatives without ionizing radiation. These latter imaging modalities are still limited by the absence of larger validation studies and standardized protocols, particularly also for young infants and newborns.

FIGURE 1.

Micturating cystourethrogram demonstrating left sided vesicoureteral reflux with contrast agent descending from the bladder through the left ureter to fill the renal pelvis and calices. No contrast flow into the right ureter.

3.2. CT with or without contrast

3.2.1. Technical notes

CT is based on the concept of x‐rays emitted from a rotating x‐ray tube passing through the body and received by a detector placed diametrically opposite, then reconstructed into a gray‐scale image based on the density of structures within the object. The ionizing radiation used is typically around 2–3 mSv for an abdominal and pelvic CT scan, the equivalent to 200–300 chest x‐rays (Huda & Vance, 2007). Multidetector CT with multiple rows of detectors can acquire multiple separate imaging layers per rotation in a much shorter time (<10 s) and with less radiation in protocols optimized for small children (Nievelstein et al., 2010). Thinner slices of 0.6–0.75 mm provide excellent image resolution, although images of diagnostic quality with a certain level of image noise should be prioritized over maximal resolution to reduce radiation exposure. Post‐processing of images allows for 3D reconstruction of complex anatomical structures. The standard intravenous contrast medium for CT imaging is iodinated, non‐ionic and low‐osmolar, ideally administered by power injector instead of hand injection, although care needs to be taken of the limited maximum flow rates of the small cannulas used in neonates (24 or even 26 Gauge). The delay between injection and imaging procedures varies depending on age and size of the infants and on indication (e.g., arterial phase CT vs. excretory study), and can be difficult to predict despite improved protocols (Nievelstein et al., 2010). Kinetic models for different contrast medium derive estimates of renal function, but also cortical and medullary volumes from contrast CTs (Grenier et al., 2011).

3.2.2. CT indications in newborns

CT is rarely used in the neonatal population and should never be the first choice for imaging kidney and urogenital structures. The indications for CT scans in term and preterm born infants need to be very carefully weighed against the significant radiation exposure, particularly in the very low birth weight infants or sick newborns in the NICU who undergo multiple ionizing investigations during their birth admission and beyond, often with larger than required radiation doses (Hogan et al., 2018; Su et al., 2022). Despite all the technical advances in speed and resolution of CTs, sedation is still commonly required for the neonatal patient to obtain motion‐free images. With the increasing use of MRI, very few indications for conventional and contrast CT remain. Occasionally, contrast CT combined with MCUG can reconstruct the entire renal and urogenital tract where MRI does not provide sufficient enough spatial resolution in very small patients with complex malformation (Hiorns, 2011). Rare case reports found contrast CTs to detect malformations or poorly functioning or small dysplastic kidneys where ultrasound, IVU or DMSA scintigraphies failed to locate them (Grenier et al., 2011; Pantuck et al., 1996).

3.3. Nuclear medicine imaging or scintigraphy

3.3.1. Technical notes

For nuclear medicine diagnostic studies, radioactively labeled tracer molecules (radiopharmaceuticals) are injected intravenously (or more rarely ingested or inhaled) to travel and bind to specific target cells. The radiation emitted from the binding site is then captured by gamma cameras over a specified time period with or without a delay after tracer administration, either in planar or tomographic/slice images (single photon‐emission CT SPECT). The radiation exposure from scintigraphies ranges in the equivalent of approximately 1–2.5 mSv, depending on the activity of the radioactive tracer and time from injection to excretion (De la Vaissiere et al., n.d.). Outside of the radiation exposure, the disadvantages of nuclear medicine scans lie in the limited anatomical details and spatial resolution, even in SPECT imaging. For neonates, scintigraphies require transport to nuclear medicine facilities which are frequently in remote locations of hospitals, thus include challenges of thermoregulation during long acquisition times.

3.3.2. Static and dynamic renal scintigraphies

Technetium labeled dimercaptosuccinic acid (^99mTc‐DMSA) binds to the proximal convoluted tubules after intravenous injection, with images acquired 2–4 h after administration to provide a static image of the renal parenchyma. The main indication is the detection of renal scarring (for which the SPECT DMSA has become the standard) and other focal parenchymal anomalies (Figure 2) (de Bruyn & Marks, 2008; Demirel et al., 2012). Quantitative analysis of the photon emissions also allows for calculation of differential renal function.

FIGURE 2.

Technetium labeled dimercaptosuccinic acid (^99mTc‐DMSA) scintigraphy as planar scan demonstrates two defects on the lateral and upper pole of the right kidney (arrows on posterior view in panel a and on anterior view in panel b). Single photon emission computed tomography (panel c) unveils additional defects in the left kidney (arrows) (adapted with permission from Einarsdottir et al. [2020]).

For dynamic renal scintigraphies, radioactive mercaptoacetyltriglycine (^99mTc‐MAG 3) is injected intravenously and extracted from blood into proximal tubules, from where it passes with urine through the collecting system into the bladder. Image acquisition usually takes at least 25–30 min to capture perfusion, extraction and excretion phase, often enhanced by administration of diuretics such as Frusemide. Indications include the evaluation of renal tract obstruction and calculation of differential renal function as well as perfusion studies. The dynamics of ^99mTc‐MAG 3 functional scans change with age: the newborn has a relatively short time to peak perfusion (T _max), which increases in young infants and slowly decreases again until 3 years of age. In contrast, the half‐life (T _1/2) is longest in newborns and becomes shorter until age three (Girotto et al., 2014). Assessment of differential renal function by ^99mTc‐MAG 3 is similar to ^99mTc‐DMSA scan in pediatric patients (Demirel et al., 2012). The kidney length measured during the parenchymal perfusion phase of the nuclear medicine scan is comparable to the length measured by ultrasound (Girotto et al., 2014), but this would never be the primary indication for nuclear medicine imaging. ^99mTc‐diethylene triamine pentaacetate (DTPA) is only excreted via glomerular filtration and hence ideal for direct assessment of GFR. However, ^99mTc‐DTPA distributes freely into the relatively large extravascular space of a newborn and has lower renal extraction rate compared to the better suited ^99mTc‐MAG3 in this group of patients (de Bruyn & Marks, 2008). Both static and dynamic scintigraphies can be performed soon after birth and are feasible in preterm infants, but the quality of the imaging improves with renal maturation over time (de Bruyn & Marks, 2008).

3.3.3. Radionuclide cystography

Radionuclide cystography applies the same principle as MCUG, but a radiopharmaceutical (often ^99mTc‐DTPA) is instilled into the bladder rather than an iodized contrast agent. The modality is useful to assess VUR or bladder obstruction with much lower radiation dose than MCUG, but at the cost of loss of anatomical details, particularly in the area of the urethra. It is therefore more commonly used in girls in whom anterior urethral valves are an absolute rarity (de Bruyn & Marks, 2008).

3.4. Ultrasound

3.4.1. Technical notes

Ultrasound imaging, or sonography, produces images of anatomical structures noninvasively and without ionizing radiation by emission and reception of sound waves into and from tissue. Ultrasound probes, so called transducers, send beams of sound waves into the tissue that then reflect on structural boundaries (e.g., between fluid and soft tissue, between soft tissue and bone, between dense and less dense tissue structures) and travel back to the transducer at different speeds, to be transformed into a two‐dimensional gray‐scale image. These typical sonograms are referred to a B‐mode (brightness mode) images where each bright dot represents an ultrasound echo.

The shape and size of transducer footprint and the frequency of the emitted sound waves determine the depth and details of the imaged organ structures. Linear probes emit parallel high frequency ultrasound waves and produce images of superficial structures of similar high resolution across the wave front. Curvilinear transducers emit a sector shaped low frequency beam which fans wider while traveling deeper through tissue, creating an image wider than the actual footprint of the probe. The dispersion of the waves at the periphery results in poorer resolution and poorer image quality of the lateral and deep structures. Phased (or sector) array transducers benefit from a compromise of qualities of linear and curvilinear probes: they have small footprints with a narrow beam sector of low frequency sound waves, reaching more depth than linear probes, but also better lateral resolution than bigger curvilinear probes. High frame rates enable excellent temporal resolution for fast moving structures such as the heart.

Ultrasound imaging is widely available for diagnostic purposes in newborns, frequently used as bedside point of care imaging tool by clinicians in the NICU (Kluckow, 2014). It is feasible even in unwell and extremely preterm neonates and does not require sedation. Most structures of interest in neonates are relatively superficial and small, so that many ultrasound studies can be obtained with excellent image details using a linear probe with high frequency.

3.4.2. B mode ultrasound

Estimation of kidney size is a frequent indication of neonatal renal ultrasound for a variety of reasons. Kidney length is proportional to birth weight and gestational age in newborns, but also variable between individuals depending on body size and sex (Figure 3a) (Al Salmi et al., 2021). While kidney length appears to be an easily reproduced and accurately estimated, the small organ size in young children and operator dependence can introduce interobserver variability in the order of 2 years of kidney growth (Al Salmi et al., 2021). VUCG remains the gold standard for diagnosing VUR, but incorporating sonographic evidence of abnormal renal length and parenchymal thinning increases the diagnostic sensitivity for clinically relevant higher degree of VUR (Muensterer, 2002). Calculation of kidney volume requires more measurements taken in different planes, consequently greater risk of inaccuracy, but gives a better estimate for actual kidney size and growth than just renal length because of additional information on thickness and length (Al Salmi et al., 2021). There is debate how well estimation of kidney volumes (as a surrogate marker for kidney weight) can be used to infer total nephron numbers in neonates (Al Salmi et al., 2021; Kent et al., 2010). Validation of sonographic kidney volume as surrogate measure for oligonephropathy could single out those preterm born patients at higher risk for poor kidney health later in life (Kandasamy et al., 2013). Some studies propose the measurement of renal parenchymal thickness (from capsule through to cortex and medulla to the pyramidal apex adjacent to the renal pelvis) for assessment of the nephrogenic zone and, indirectly, total nephron mass. Preterm infants at the corrected age of 32 weeks were found to have narrower thickness compared to measurements in utero at the same gestational age, consistent with potential lower nephron numbers after experiencing limited ex‐utero nephrogenesis after preterm birth. Similar thickness at 37 weeks may indicate compensatory glomerular hypertrophy closer to term (Brennan et al., 2022).

FIGURE 3.

(a) B‐mode ultrasound image of a kidney in a newborn infant. The Doppler sonography in panels (b, c, and d) overlay the gray‐scale B‐mode image: (b) Power Doppler ultrasound displaying the total blood flow in the kidney, with continuity throughout the parenchyma. (c) Color Doppler ultrasound of renal artery and vein at the hilum of the kidney. (d) Microvascular flow imaging (here: Superb Microvascular Imaging) displaying blood flow from the interlobar artery into the small cortical arteries close to the renal capsule.

Assessment of the renal parenchyma includes the appraisal of cortical and medullary echogenicity. The cortex should be less echogenic than the adjacent liver or spleen. The medulla is relatively hypoechogenic to the cortex in newborns, resulting in a good corticomedullary contrast (Gothi & Raj, 2022). Congenital bilateral cortical hyperechogenicity raises suspicion of hepatorenal fibrocystic diseases, of which the spectrum of polycystic kidney diseases comprises the largest group (Avni & Hall, 2010). Serial high‐resolution ultrasound can display the evolution of debris collection in the pyramidal tips collection after birth, some of which can be severe enough to cause obstruction to urinary flow into the collecting system in the first days of life (Gothi & Raj, 2022).

The most common indications for renal ultrasound in the neonatal period are follow up of antenatally detected urogenital abnormalities, mainly urinary tract dilatation which is found in 1%–5% of antenatal ultrasound examinations (Nguyen et al., 2010). The latter is variably referred to as hydronephrosis, pyelectasis, pelviectasis, or pelvicalyceal dilatation, and evaluated by measuring the widest intrarenal diameter of the renal pelvis in an anterior–posterior aspect on the transverse B‐mode image of the kidney. Additionally, dilatation of calices, parenchymal thickness and echogenicity are evaluated. Recommendation for classification and nomenclature aim at standardizing the assessment of a condition that is often benign and transient, but can also be a symptom of VUR, pelviureteric junction obstruction or other urological malformations (Chow et al., 2017). Of note is that a non‐dilated appearing drainage system does not rule out higher degree vesico‐ureteral reflux, for which VUCG remains the diagnostic standard (Nelson et al., 2014).

3.4.3. 3D/4D ultrasound

For 3D ultrasound, positional information from position sensors attached to the transducer is used to reconstruct a set of 2D ultrasound images into static spatial impressions of anatomical structures. Newer technologies with motors integrated into transducer heads perform sweeps of predefined areas and reconstruct 3D images in real‐time during scanning (referred to as 4D ultrasound). The rendered 3D images offer good impressions of surfaces of oblique or irregularly shaped objects, for example a scarred kidney surface or a tortuous course of the drainage system. Volume of 3D shapes can be calculated post‐hoc, or electronic scalpels applied to slice organs in the desired plane (Huang & Zeng, 2017).

Estimation of kidney volumes by 3D ultrasound has been found to provide accurate results compared to MRI or CT, and more reliably than 2D ultrasound (Fritz et al., 2003; Kent et al., 2010). This is particularly true for irregularly shaped kidney surfaces and in hydronephrotic kidneys, where the dilated collecting system can be subtracted in the 3D rendering process (Riccabona, 2011). Split volumes of renal parenchymal volume on 3D ultrasound correlate well with findings on ^99mTc‐DMSA scintigraphy or MRU. Assessment of parenchymal echogenicity and cortico‐medullary differentiation remains superior in B‐mode images compared to 3D ultrasound because the resolution and quality are consistently better on 2D images. Cysts are easily visualized and can be followed longitudinally with good standardization (Riccabona et al., 2003). The technique lends itself to visualization of hydronephrosis (Figure 4). Rendered views of complex malformation of the drainage system are equal to the views obtained by MRI or IVU and are clearly a strength of the noninvasive 3D imaging technique. Similarly, the rendering of the internal bladder surface is akin to a “virtual cystoscopy” with 4D ultrasound (Riccabona et al., 2003).

FIGURE 4.

A 3D ultrasound rendered view of a megaureter in a neonate with additional uretero‐pelvic junction obstruction due to a crossing aberrant vessel (zoomed image on right) (reprinted with permission from Riccabona et al. [2003]).

The main limitations of 3D ultrasound remain the lower resolution of reconstructed planes compared to B‐mode images. The quality of 3D images relies directly on the quality of the 2D sonography and may be further impaired by motion artifacts which distort the rendering process (Riccabona, 2011).

3D and 4D ultrasound, while routinely used on obstetric imaging, has so far not been widely implemented into neonatal scanning protocols, and if at all mostly in brain ultrasounds (Kishimoto et al., 2013). There are a few reports of renal and urogenital 3D imaging in patients of neonatal age, but larger scale studies are still lacking. As with all ultrasound modalities, the potential of noninvasive longitudinal studies is an attractive argument to implement the technique more widely, for example to track pelvicalyceal dilatation and kidney growth.

3.4.4. Doppler sonography

Doppler sonography takes advantage of the Doppler shift effect where frequencies of sound waves change between transmission into and return from the tissue due to the reflection off of moving objects (such as blood cells flowing in blood vessels). In color Doppler ultrasound, the speed and direction of the movement of blood is expressed in a scale of colors in real time—a distinct color for each direction of flow (generally blue for flow traveling away from the transducer, and red for the flow traveling toward the transducer); shades of the two colors denote higher or lower mean velocities. The method is excellent for demonstrating the vascular anatomy of an organ and the direction of blood flow (Figure 3c). Spectral Doppler sonography, also called Duplex sonography, provides the spectral waveform of the blood flow displayed in the color Doppler image. For pulse wave spectral Doppler, the transducer sends sound wave pulses at a certain pulse rate and receives the reflected frequencies of each moving blood cell from a specified area (where a gate is placed within a blood vessel). The spectrum of frequencies is averaged, converted into velocities and displayed as Doppler waveforms. The angle of insonation should be as parallel as possible to the blood flow, otherwise the flow velocity will be underestimated (albeit newer US device are better able to provide accurate flow data at angles up to 60°) (Silveira et al., 2000).

The first days and weeks after birth are characterized by rapidly falling peripheral vascular resistance during the newborn's circulatory transition from the intrauterine, placenta‐dependent environment to the extrauterine setting. Renal blood flow increases simultaneously from around 5% of cardiac output in utero to 15%–20% at 6 weeks of age (Sulemanji & Vakili, 2013). Color Doppler and Duplex studies of the kidney vasculature are frequently used imaging modalities in term and preterm infants during this time of circulatory transition. Early studies of renal Doppler ultrasound demonstrate the increasing renal blood flow velocities in the first week of life (Cleary et al., 1996), with a pattern of higher velocities in central hilar arteries and lower flow velocities in peripheral (interlobar or arcuate) arteries (Stavel et al., 2004). Renal blood flow is significantly lower in small for gestational age infants compared to age and weight matched controls (Kempley et al., 1993). Resistive indices (RI) reflecting the peripheral vascular resistance are consistently higher in newborns than RIs in adult (where the threshold for normal resistance is <0.7) (Cleary et al., 1996; Yildirim et al., 2005), and decrease along the vascular tree into the parenchymal arteries (Stavel et al., 2004). The presence of hemodynamically significant PDA in preterm infants produce a typical spectral waveform pattern in the renal artery Doppler: very low to often absent or reversed diastolic flow represent the diastolic steal through a left‐to‐right shunt across the PDA (Bomelburg & Jorch, 1989). Renal vein thrombosis is the most common thrombotic event in neonates and presents with characteristic ultrasound B‐mode and Doppler findings. Initially, these do not necessarily show the absence of flow in the renal veins. Typically, first sings are swelling of the kidney, alteration of corticomedullary echogenicity and differentiation and reversed end‐diastolic flow in the renal artery on the spectral Doppler waveform. Flow interruption in the renal vein follows in the days after the initial event (Mikolajczak et al., 2018).

In Power Doppler ultrasound, the amplitude or magnitude of the Doppler signals is color coded to demonstrate the total blood volume in a tissue rather than speed or direction of flow in individual blood vessels. The modality is approximately three times more sensitive than conventional color Doppler. Hence, blood flow into smaller blood vessels can be visualized (Figure 3b) (Aziz et al., 2022; Yoo et al., 2020). Power Doppler is useful to assess impaired cortical perfusion globally (e.g., in renal vein thrombosis; Mikolajczak et al., 2018) or in focal perfusion defects such as infarcts (Gordon & Riccabona, 2003), although has to date not routinely been used nor studied in neonates.

Traditionally, color and Power Doppler are limited by the low sensitivity for low‐velocity flow in blood vessels smaller than 1–2 mm. Microvascular flow imaging (called Superb Microvascular Imaging [SMI], Slow Flow, Microvascular Imaging, or MV‐Flow by different commercial manufacturers) offers visualization of ultra‐low‐flow Doppler signals in very small blood vessels using various declutter algorithms that suppress the noise caused by tissue motion and other artifacts. Generally, microvascular flow imaging is displayed in a similar way to color Doppler imaging, overlaying B‐mode images (Figure 3d) (Aziz et al., 2022). Ultrafast plane wave ultrasound transmits tilted plane waves into tissue which then backscatter and are summed up to produce high resolution images at very high frame rates. Color Doppler and spectral Doppler waveforms can be acquired simultaneously during a single acquisition, visualizing up to three waveforms and extraordinarily detailed images of vascularity at the same time (Correas et al., 2016).

In adult kidneys, microvascular flow imaging reliably detects infarcts, cortical necrosis, focal ischemia or tumors, as well as kidney graft rejection (Correas et al., 2016). The high resolution of these microvascular flow studies come at the cost of image degradation in deep tissue (Aziz et al., 2022). However, given the smallness of neonatal patients with organ structures directly under the surface, the micro‐Doppler techniques lend themselves for remarkable displays of small caliber vasculature beds in this patient category. In newborns, microvascular flow studies have been applied to cranial ultrasound with documentation of gestational maturation of superficial microvasculature (Barletta et al., 2021; Goeral et al., 2019). There are no studies to date that used the technique in neonatal kidneys, but the comparison of microvascular with conventional color Doppler ultrasound in pediatric patients showed superior display of parenchymal arteries to the interlobar level and clear definition of the medullar pyramids against the cortex. In acute pyelonephritis, microvascular ultrasound delineated areas of hypoperfusion much more clearly, as well as stretching of interlobar blood vessels as an indirect sign of edema (Yoo et al., 2020). Imaging software facilitates quantitative assessment of renocortical vascularity on SMI ultrasound by applying color pixel identification and subtracting color signals from the areas of interest in the kidney cortex (Gao et al., 2019). This simple image analysis algorithm on a noninvasive and radiation free modality opens the exciting possibility to quantify the development of renal microvasculature in neonates to longitudinally describe the changes over increasing gestational age.

Renal ultrafast Doppler has to date mostly been used in animal models, where unilateral ischemia–reperfusion injury leads to an obvious and histologically confirmed rarefication and increased tortuosity of the renal vasculature (Chen et al., 2020). The technique has shown promise in human neonates as a functional application, demonstrating changes in cerebral blood flows during different stages of sleep cycle and during seizures (Demene et al., 2019), but has not been applied to similar studies of kidneys.

3.4.5. Contrast enhanced ultrasound

Contrast‐enhanced ultrasound (CEUS) uses nonionizing microbubble contrast agents consisting of high‐molecular weight gas within a protein, lipid or polymer shell. After intravenous injection, the microbubbles remain intravascularly (i.e., do not enhance the drainage system like CT or MRI contrast agents, unless retrogradely instilled for diagnosis of VUR; Gordon & Riccabona, 2003). With a diameter of 2–6 μm they distribute into small capillary vascular beds, where ultrasound waves provoke oscillations of the bubbles and returning of frequencies distinguishable from the tissue clutter. With a half‐life of 5–7 min, the microbubble shell then degrades and is excreted through the liver, while the gas is exhaled through the lung, making it suitable for patients with impaired renal function. The resulting greyscale or color images are of excellent quality and provide detailed evaluation of microvasculature of the renal blood vessels as small as 40 μm. CEUS can be used to estimate renal blood flow and fractional blood volume of the microvasculature. Tissue perfusion is measured by analyzing the replenishment kinetics of the volume of microbubbles after their destruction by high‐power insonation (Grenier et al., 2011). The modality is superior to CT and MRI mainly because of the high temporal resolution, allowing for real‐time imaging. In combination with microvascular flow imaging, CEUS augments the display of the renal microvasculature by temporal summation of the microbubble trajectory (i.e., keeping the echo of the microbubble over time) (Correas et al., 2016). The modality has not yet been explored in young children or newborn infants, and there are no safety data for microbubble contrast in this patient group. Outside the main indications in adults (evaluation of masses and cysts or parenchymal ischemia or longitudinal disease progression; Cokkinos et al., 2013; Wang et al., 2018). CEUS could be useful to track renocortical vascular development and/or quantify renal perfusion over different gestational ages. To achieve this, standard parameters for CEUS microperfusion studies need to be established first (Figure 5) (Zhang et al., 2022).

FIGURE 5.

Contrast enhanced ultrasound of a kidney in a healthy child (top panel) and in a pediatric patient with chronic kidney disease (bottom panel). Three regions of interest (ROI as yellow, green, red box) are placed in the cortex, then quantitative parameters of wash‐in blood flow are generated for each ROI. Time to intensity curve displayed to the right of the ultrasound image, with the following parameters displayed: TtPk = time to peak intensity, interval between start of enhancement to peak contrast intensity, A = plateau value as an estimate of the regional blood flow, B = baseline, k = slope of the ascending intensity curve. All parameters are decreased in the diseased organ (reprinted with permission from Zhang et al. [2022]).

3.4.6. Ultrasound elastography

Ultrasound elastography is based on the principle that a compressing force applied to tissue causes small displacements which in turn are then measured as shear waves in relation to the applied force. The results are expressed quantitatively or as color‐coded maps over native B‐mode images. They give an estimation of stiffness of the underlying tissue: the stiffer a tissue is, the faster shear waves will spread through it. The applied force can be external (by compression via the ultrasound transducer) or internal by acoustic radiation force impulse (ARFI). Elastography is most widely used and tested in liver fibrosis when changes in the tissue architecture of parenchyma have not yet become visible in B‐mode ultrasound (Ferraioli & Roccarina, 2022). In theory, glomerular or interstitial fibrosis in the kidney can develop well before it affects the measured global renal function, but could be found and quantified earlier by elastography. Kidney elastography has mostly used ARFI techniques, as constant and reproducible external pressure is very difficult to achieve through the whole depth of the organ. The estimation of kidney stiffness is subject to many confounding factors, namely by renal perfusion, age and the lack of a uniform approach to measurements. Interestingly, stiffness is inversely proportional with age, albeit data for newborns are lacking. Even minimal compression through the transducer interferes with the measurements in the smaller patients who are therefore not ideal subjects (Correas et al., 2016). A single study assessed renal elastography in former preterm children at 10 years of age and found abnormal (higher or lower) values compared to term born children, possibly related to underlying prematurity‐induced histological changes that affect stiffness and deformity properties (Zaffanello et al., 2015). The modality could be promising to pick up developing renal fibrosis early, but larger studies are required to establish normal values and consistent protocols in young children and newborns.

3.5. Magnetic resonance imaging

3.5.1. Technical notes

The technique of MRI is based on the principle that hydrogen protons in water molecules align when exposed to a strong magnetic field. During the imaging procedure, protons are forced out of alignment by pulses of radiofrequency current, then relax back into the previous equilibrium within the magnetic field once the emitting current stops. The signal received by the coil from each voxel (a volume element in a three‐dimensional space) reconstructs an image depending on the time protons require to relax and the energy released during realignment. The time taken for the protons to relax and realign is measured in two different ways: T1 relaxation time is the time protons take to longitudinally relax into their original magnetic vector; T2 relaxation time measures the time for the axial spin of protons to return to the resting state. Consequently, in T1‐weighted MRI sequences, the contrast between anatomical structures corresponds to the differences in T1 relaxation times, and vice versa, T2 properties of a tissue dominate the contrast and brightness of an image in T2‐weighted sequences.

Contrast agents, most commonly gadolinium based chelates, act as powerful local magnets that force protons to relax instantaneously, therefore emitting a high intensity signal. Intravenous injection of gadolinium visualizes vascular structures and traces blood supply in organ structures. There is a dearth of safety data on gadolinium‐based contrast agents in neonates (Fraum et al., 2017). Reassuringly, the newest evidence has defused concerns of gadolinium accumulation in fetal tissue and associated risk of intrauterine death or neonatal morbidities after contrast MRI in pregnancy (Winterstein et al., 2022). The risk of nephrogenic systemic fibrosis in children appears very low (Nardone et al., 2014). But specific consideration or recommendations for the use of iv gadolinium in neonates with their naturally low GFR are lacking.

The use of MRI in preterm and term neonates, mainly brain imaging, has increased rapidly over the past decade as facilities have become more readily available (Arthurs et al., 2012). Higher imaging speed, improved spatial resolution, motion correction, and more refined signal‐to‐noise ratio have increased the utility of MRI to visualize the structures of other organs, notably those of the kidneys (Hillenbrand & Huisman, 2012). Clinicians value the nonionizing method for its excellent imaging quality, but this is offset by the requirement of sedation and transport to scanners which often are remote from the NICU and present challenges for thermoregulation, monitoring and noise exposure for the vulnerable patient group (Arthurs et al., 2012). Financial constraints frequently result in a lack of specific equipment such as bespoke neonatal coils, instead adult equipment (such as knee coils) is “made do” for use in newborn imaging (Arthurs et al., 2012). Specifically designed MRI systems within or close to the NICU have been shown to deliver imaging quality in neonates similar to state‐of‐the‐art adult scanner (Tkach et al., 2012).

3.5.2. Kidney MRI, MRU, fMRU

One of the earliest and still most common applications of renal MRI in young infants is the assessment of hydronephrosis and complex kidney or urinary tract malformations (Hiorns, 2011). Heavily T2‐weighted images highlight areas of high‐water content, hence provide excellent endogenous contrast of dilated structures of the collecting system, regardless of renal function (Figure 6a). MRU uses intravenous gadolinium contrast to further enhance the delineation of the fluid spaces of the urinary tract, frequently also used in the preoperative planning and postoperative follow up of pyeloplasty (Kirsch et al., 2006; Riccabona, 2008; Rodriguez et al., 2001). MRI and MRU can deliver at least equal, if not better images of duplex kidneys and associated complications of the drainage system compared to the combination of ultrasound, VCUG, IVUs and nuclear medicine scans, with good interobserver agreement on anatomical details (Figure 7) (Avni et al., 2001; Rodriguez et al., 2001).

FIGURE 6.

Coronal magnetic resonance imaging sequences of a 3‐year old patient. (a) T2‐weighed MR images show a normal kidney with a cyst (arrowhead). The parenchyma is of normal signal intensity with adequate corticomedullary differentiation. No calyceal diverticulum was identified. (b) Fractional anisotropy shows pyramids in light gray (arrow) and cortex in dark gray (dotted arrow). (c) Diffusion tensor imaging by parenchymal tractography shows tracks from different regions of the kidneys and how they coalesce into well‐defined geometric structures (arrows). (d) Overlay of coronal‐parenchymal tractography and fractional anisotropy map shows that the geometric structures coincide with the medullary pyramids (reprinted with permission from Jaimes et al. [2014]).

FIGURE 7.

T1‐weighted post‐contrast images show pelviureteric junction obstruction on the right side (b) and hydroureteronephrosis on the left side (a). Contrast passes through kidney (renal phase, left panel), then passes through collecting system (excretion phase, right panel) (reprinted with permission from Sharma et al. [2016]).

T1‐weighted MR images provide good quality images of the solid organ components of the kidney (Figure 7). Kidney volumes can be reliably assessed from MRI images. However, they systematically underestimate the volume compared to postmortem assessment, likely due to the inaccuracy of tracing on MR images for volumetric calculations (Prodhomme et al., 2012). Contrast enhanced MRI yields more details of the renal parenchyma, for example of the cortico‐medullary differentiation, than non‐contrast MRI and differentiate better between foci of acute inflammation and scarring in acute pyelonephritis than static DMSA nuclear medicine scan (Kirsch et al., 2006). An important indication for MRI with intravenous contrast is functional MR urography (fMRU) where the uptake, excretion and draining of gadolinium is visualized sequentially, providing estimates for renal transit time and differential renal function. The functional studies correlate well with those obtained with MAG3 renal scintigraphy (Kirsch et al., 2006). Analyzing the excretion of the contrast agent over time transforms the morphological information obtained during the MRI into estimates of GFR (Jaimes et al., 2014; Kirsch et al., 2006). The difficulty of GFR estimation from contrast studies in newborns (not just fMRU, but for any imaging modality) is the dynamic renal function, with rapidly changing blood flow and secretory and resorptive capacity over the first days and weeks of life (Smeets et al., 2022). Cystatin C might be a more suitable indirect marker than creatinine to calculate eGFR after birth (Filler et al., 2021). But renal function measurements from contrast studies are commonly validated against eGFR from creatinine (Buchanan et al., 2020) and, even for this method, there is no standard for the first weeks of life or in the setting of the developmental changes after preterm birth.

3.5.3. Other MR sequences

The combination of specific MRI sequences with application of software‐based calculations generate MR images that take advantage of distinct characteristics and movements of tissue components. Diffusion‐weighted images (DWI) measure the random Brownian movement of water molecules within a voxel of tissue and exploit the differences in these movements to generate imaging contrast. Tissue components with higher cellularity, or injured tissue with swelling, exhibit lower diffusion, that is, less movement of water and therefore less signal intensity. From DWI, the magnitude of water diffusion can be quantified and expressed as apparent diffusion coefficient (ADC).

Diffusion tensor imaging (DTI) and fractional anisotropy (FA) maps benefit from water movement in the highly organized radial arrangement of tubules and collecting ducts within the renal medulla. FA images provide information on microstructural integrity of the cortex and medulla (Figure 6b,d). Through 3D modeling technique (tractography) DTI provides detailed images of the spatial architecture of tracks throughout the renal parenchyma, including the organization of the medulla into pyramids (Figure 6c,d). Importantly, the strict radial organization of the tracts is lost in kidneys affected by pelvicalyceal dilatation (Jaimes et al., 2014).

The majority of the MR modalities described above have either only been studied in adult subjects or animal models or on ex vivo kidneys altogether. Very few studies have explored basic MR techniques on the kidneys and urogenitary tract of term or preterm newborns. FA maps correlate with renal function when combined with fMRU in children (Jaimes et al., 2014) and potentially indicate early kidney damage prior to becoming clinically apparent in adults (Wang et al., 2018). The example of the progression of ADC signal in fetal kidneys (Witzani et al., 2006) and over the first years of life (Jones & Grattan‐Smith, 2003; Kocyigit et al., 2014) demonstrates that signal intensity information from these types of scans cannot be extrapolated from adults to maturing infants, and that ADC does not appear to correlate with the degree of morphological abnormality or with age (Bedoya et al., 2019). Benchmarking for individual sequences at different gestational ages is paramount before meaningful inferences on normal and abnormal structural development can be drawn. For brain imaging, DTI has previously been explored in preterm and term neonates, demonstrating marked microstructural white matter changes after preterm birth (Dibble et al., 2021). These types of studies for kidney structure in this population are largely lacking at this stage.

Other MRI sequences rely on blood flow and oxygen delivery to tissue to generate functional studies from morphological information. Blood oxygen dependent (BOLD) MRI takes advantage of the paramagnetic properties of deoxygenated hemoglobin where even small increases in oxygen demand with higher proportion of deoxygenated hemoglobin lead to local magnetic field distortion in and around the blood vessels in the renal medulla. This concept has been validated in the kidneys of small animals and in humans to demonstrate medullary hypoxia (Grenier et al., 2011), but has not been explored in preterm infants who have shorter tubuli and underdeveloped medullary structures.

Arterial spin labeling (ASL) MRI uses noninvasive magnetically labeled protons in arterial blood as an endogenous tracer. The flow labeled image of the vascular tree is then compared to and subtracted from the native image, with the remaining signal proportional to the arterial perfusion of the kidney. ASL has been validated against other perfusion study modalities in humans and animal models and is deemed to produce reproducible perfusion measurements suitable for longitudinal studies (Odudu et al., 2018).

BOLD and ASL MRI are being established as part of multi‐parameter MRI, in which multiple non‐contrast dependent sequences from a single scan session demonstrate different aspects of kidney structure and function. Results from multi‐parameter MRI obtained from T1‐weighted mapping, DWI, ADC, ASL, non‐contrast MR angiogram and BOLD showed reliable and reproducible correlation compared to histological results from biopsy and GFR assessments in healthy subjects and those with chronic kidney disease. Importantly, multiple parameters together appear to provide a more complete picture of renal structure and function than the individual parameters in isolation (Buchanan et al., 2020). Changes in renal cortical perfusion in multiparameter MRI have been correlated with increasing degrees of interstitial fibrosis and interpreted as vascular rarefication in chronic fibrosis in adults (Buchanan et al., 2020). Whether this could be applied to demonstrate the degree of vascular development in the newborn kidneys remains to be studied.

3.5.4. Experimental MR modalities

All of the above MRI sequences are increasingly used in a clinical context to evaluate kidney structure and function, albeit predominantly in adults suffering from various forms of chronic kidney disease, but not to visualize structural development in newborns. Meanwhile, more MR methods are being explored ex vivo or in animal models that show great promise for visualization of renal structures. Cationic ferritin (CF) is a nanoparticle contrast agent with cationic charge which electromagnetically binds to and accumulates in the glomerular basement membrane, therefore visualizing individual glomeruli. This allows counting of total nephron numbers and the measurement of their size in 3D MRI (Figure 8a) (Charlton et al., 2021). Agreement between MRI based glomerular count and stereological methods on histological cuts are excellent, but less so for estimates of glomerular volume (Bertram et al., 2014). Whole kidney labeling with CF shows the distribution of glomeruli and cortical vasculature across the cortex (Parvin et al., 2020). CF MRI has great potential to replace currently inaccurate surrogate measurements for total nephron number, which can be heterogenous across the cortex and change over time with compensatory hypertrophy in the context of congenital nephron deficit (Charlton et al., 2021; Kent et al., 2009). However, some technical issues remain to be solved before the method becomes applicable in vivo in human subjects, for example, in damaged or not fully perfused glomeruli (Beeman et al., 2014). CF dose finding studies will have to be performed when biosafety of CF in humans has been established (Beeman et al., 2013; Charlton et al., 2016).

FIGURE 8.

(a) Cationic ferritin (CF)‐enhanced magnetic resonance imaging of an ex vivo human kidney demonstrating each glomerulus as a punctuate dark spot. (b) Blood vessels (red arrow heads) are outlined on magnified CF‐enhance magnetic resonance images for rendering of the vascular tree (c) (adapted with permission from Charlton et al. [2021]).

Other experimental MRI sequences currently explored ex vivo or in small animal models are image segmentation of combined 3D spin echo and CF enhanced MRI to render the vascular tree within the kidney in 3D (Figure 8b,c) (Charlton et al., 2021). The mapping of arterial diameters suggests that acute kidney injury leads to significant arterial vasodilation and eventually to decreased arterial density when glomeruli are lost (Parvin et al., 2020). Other applications of 3D facilitate mapping of renal pyramids, finding higher number of pyramids than previously assumed from autopsy studies. These studies also postulate that subpyramidal structures are determined by interlobar vessels, therefore representing the extent of vascularization (or, in the setting of poor kidney health, the loss thereof) (Charlton et al., 2021). All of these preliminary studies are yet to be translated into the clinical setting, leaving questions regarding their utility to demonstrate morphological structure currently unanswered.

3.6. Kidney biopsy

3.6.1. Technical notes

To obtain renal tissue for histopathological examination in vivo, a cylinder of renal parenchyma is retrieved with an appropriately sized needle through a kidney biopsy. The latest recommendations by the European Society of Pediatric Radiology task force specify optimal procedures for renal biopsy in children, including appropriate analgo‐sedation (involving an anesthetic team), positioning (best in prone), pre‐procedure ultrasound to determine optimal site (usually the lower pole of the native left kidney), and then biopsy under real‐time imaging via color Doppler ultrasound with a semi‐automated biopsy device. These recommendations state an optimal needle caliber of 18–20 G for a sample length of 12 to 22 mm in the medio‐dorsal avascular line, parallel to the main vascular plane to avoid injury to the larger blood vessels and collecting system. The tissue sample should be inspected under the microscope immediately after retrieval, as an adequate number of 10–20 glomeruli are paramount for diagnostic success, and a second pass attempted in case of insufficient sample (Riccabona et al., 2014). While these recommendations do not exclude newborns, they do not offer specific provisions for this patient group. As biopsies in newborn are rarely performed, there are no clear guidelines and experience often anecdotal. Newborn cortical thickness is only around 2‐3 mm (compared to 10 mm in older children and adults), therefore the chance of an appropriate sample with enough glomeruli is smaller. One center describes their experience with a tangential approach of 45° angle to the skin surface for younger pediatric patients under general anesthesia, yielding higher total number of glomeruli with more passes and still lower complication rates compared to the more commonly used perpendicular approach in light sedation (both methods performed with 18 G needles) (Pettit et al., 2022). The youngest patient in this cohort was a preterm infant of 31 weeks gestational age and 1.8 kg weight (personal communication of the senior author of Pettit et al. [2022]).

3.6.2. Indications for biopsy

Typical indications for renal biopsy in children include proteinuria, recurrent gross hematuria, atypical or steroid resistant nephrotic syndrome, rapidly progressive glomerulonephritis, and acute kidney injury with a duration of over 4 weeks without clear cause (Nicholson et al., 2000). None of these are common neonatal renal problems, apart from early onset chronic kidney disease of unknown origin. If at all, biopsies in newborns are considered when congenital nephrotic syndrome, Denys–Drash syndrome or ACE‐inhibitor fetopathy are suspected (Pettit et al., 2022). Many of these diseases are genetic conditions, so that the wider availability and faster turnaround time for sophisticated genetic studies including whole genome sequencing have decreased the requirement of renal biopsies in the youngest and smallest infants (Daoud et al., 2016; NICUSeq Study Group et al., 2021).

On autopsies of preterm infants and from biopsies of older preterm born children, clear indicators of nephron deficits and aberrant glomerulogenesis and subsequent hyperfiltration are described, resulting in larger glomeruli with sclerotic changes (Koike et al., 2017; Sutherland et al., 2011). However, renal biopsy is an unsuitable diagnostic procedure for the assessment of structural or morphological changes in the kidneys that are due to prematurity or critical illness in the newborn period. These maturational deficits have no short‐term consequences for the management of the preterm neonate; given the invasiveness of histopathological examination, kidney biopsies cannot be used for longitudinal follow up. Furthermore, methods to estimate total nephron number in vivo are lacking. The number of glomeruli on a biopsy specimen can be used for inference of total nephron mass, but for a true estimate the total cortical mass would have to be known. Some studies have combined the glomerular density from biopsy specimens with CT imaging or cortical volumes estimated on MRI to estimate total nephron numbers, but the methods are not established in clinical practice (Charlton & Abitbol, 2017).

4. CONCLUSIONS

Since the earliest attempts of visualizing the kidney by IVUs over 60 years ago, and the introduction of ultrasound imaging around the same time, a multitude of imaging modalities have been developed to reflect structural morphology of the kidneys (Table 1). Many of them are still predominantly studied in adults with evolving or established chronic kidney disease or renal masses. Where they are used in young children or infants, the indication is mostly the investigation of congenital anomalies or antenatally diagnosed urinary tract dilatation. To date, very limited attention has been directed toward methods to study the kidneys after preterm birth, even though the abrupt interruption, then aberrant resumption of the renal development in preterm infants is considered pivotal in the pathogenesis of kidney and cardiovascular disease later in life. Such methods should ideally be noninvasive, without (or with only very minimal) radiation exposure and readily available, because they are performed to document organ development and not diagnostic procedures for malformations or for guidance of treatment (at least at this stage). Renal biopsies remain anecdotal in newborns and are too invasive and fraught with risks for the purpose of observing structural development. X‐rays with or without contrast, fluoroscopy and nuclear medicine imaging do not provide sufficient details of kidney structures. They are generally easily accessible, but, like CT scans (which give more structural details), involve relevant ionizing radiation exposure. Exceptional advances in sonographic methods have seen the introduction of a broad new range of ultrasound techniques to visualize renal morphology. Outside of the vastly improved resolution on B‐mode images, Doppler modes that quantify blood flow to and through the kidneys (Power Doppler) or demonstrate microvascularity down to the capillary size of blood vessels (microvascular Doppler flow imaging) have opened new possibilities of visualizing previously unseen vascular structures. Ultrasound imaging is established in NICU and can safely be performed bedside in the smallest and youngest infants, therefore offering uncomplicated longitudinal observations. The increasing use of ultrasound as a point of care imaging modality by the treating neonatologists or nephrologists will only add to the future potential of this method for assessing the effects of preterm birth on kidney development. CEUS offers similar details to Doppler sonography, but requires intravenous access and careful timing of administration of the microbubble contrast agent for which safety in newborns has not conclusively been established. MRI has progressed rapidly with the introduction of many new sequences, each focusing on different structural components and/or functional aspects of the kidneys. Endogenous contrast within the kidneys displays structures on MRI with remarkable detail; post‐processing renders images of morphological features in unprecedented clarity. Despite great promises for structural images pertinent in the context of prematurity, MRI remains problematic for patients in the NICU, as imaging takes place outside the NICU and requires sedation during the lengthy image acquisition. For these reasons, MRI is not feasible for longitudinal documentation of structural changes. Many of the recent MR techniques have never been explored in this young patient group, while some remain in the experimental stage and are yet to be introduced into clinical practice.

TABLE 1.

Methods of visualizing kidney structure and morphology in newborn infants.

Method	Kidney size/volume	Cortex/medulla	Vasculature	Drainage system	Function	Radiation exposure	Useability in neonates
Biopsy	−/−	(+) Often only cortex material obtained	(−) Only if cortical microvasculature contained in biopsy material	−	−	−	Not routinely done in small neonates, requires sedation
Native x‐ray	+/−	−	−	−	−	+	Readily available and fast, very limited diagnostic information
Intravenous pyelogram	−/−	−	−	+	−	+	Redundant, requires iv access
Micturating cystourethrogram	−/−	−	−	+	−	+	Not mobile, requires transport outside NICU
CT
Native	+/+	(+)	−	−	−	++	Most indications for CT have nonionizing alternative diagnostic methods. Contrast requires iv access. Short scanning time. Requires transport outside NICU and sedation
With contrast	+/+	(+)	(+) (macrovasculature)	+	(+)
Nuclear medicine scan
MAG3	+/−	−	−	+	+	+	Requires iv access and transport outside NICU. Long acquisition times.
DMSA	+/−	− (parenchyma)	−	−	+	+
Radionuclide cystogram	−/−	−	−	+	−	+
Ultrasound
B‐mode	+/+	+	(+) walls of larger vessels	+	−	−	Readily available bedside. Well tolerated by young and sick newborns
3D/4D	+/+	+	(+)	+	−	−
Contrast enhanced ultrasound	−/−	+	+	−	+ (perfusion, relative blood flow)	−	Bedside, no transport required, requires Iv access, safety of microbubble contrast agents not established in neonates
Vascular Doppler
Doppler	−/−	−	+	−	−	−	Bedside, no transport required, no iv access required.
Duplex/spectral Doppler	−/−	−	+ (flow velocities and wave forms)	−	−	−
Power Doppler	−/−	−	+	−	−	−
Microvascular flow Doppler	−/−	−	+	−	−	−
MRI
No contrast	+/+	+ (better for T2)	(−) (vessel walls)	+	−	−	Requires transport outside NICU, sedation, iv access, long scanning times
With gadolinium‐based contrast	+/+	+	(+) macrovasculature yes, microvasculature no	+	+	−
fMRU	+/+	−	−	+	+ (renal transit time and differential renal function, GFR	−	Requires transport outside NICU, sedation, iv access No standards for signal intensities in newborns, long scanning times.
DTI tractography	−	+	−	−	−	−	Requires transport outside NICU, sedation, iv access, long scanning times.
BOLD MRI	−	+ (differentiating cortex and medulla through differential oxygen supply)	−	−	+ tissue perfusion particularly in medulla	−
Multi‐parameter MRI	+	+	(+)	−	+ (renal perfusion, renal artery blood flow)	−	Not tested in neonates, limitations as all other MRI modalities
MRI with cationic ferritin contrast	−	+ (glomerular number and volume in cortex)	−	−	−	−	Not for clinical application currently, as not tested in humans

In summary, recent advances in ultrasound technique offer noninvasive, safe and nonionizing imaging of morphological details in kidneys of preterm and term neonates which allow for the longitudinal observation of development. MRI modalities promise exciting potential for unprecedented anatomical details, but require more studies in neonates including establishment of standards prior to introduction into clinical practice.

CONFLICT OF INTEREST STATEMENT

Dr Staub has no financial or other conflicts of interest to declare.

ETHICS STATEMENT

No application to the local or higher ethics committee was required for this literature review.

PERMISSION STATEMENT

Parents or guardians have provided written permission for the radiological images of their infants to be used in a de‐identified fashion in this review. Copyright permission has been obtained, where applicable, for the reproduction of figures from other articles. Figures from open access articles are referenced in accordance with the applicable guidelines of the publisher.

APPENDIX A. Search terms

Embase

Latest search 20/09/2022

1 exp newborn/

2 exp prematurity/ or

3 preterm infant.mp.

4 1 or 2 or 3

5 exp kidney

6 exp kidney structure/ or structure.mp

7 exp morphology/

8 5 and 7

9 6 or 8

10 exp kidney biopsy/

11 exp autopsy/

12 ex histology/

13 10 or 11 or 12

14 exp imaging/

15 exp X‐ray/

16 exp ultrasound/

17 exp duplex Doppler ultrasonography/

18 exp nuclear medicine/

19 magnetic resonance.mp or exp nuclear magnetic resonance/

20 exp computer assisted tomography

21 exp scintigraphy

22 exp fluoroscopy

23 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22

24 13 or 23

25 4 and 5 and 9 and 24

Medline

Latest search 20/09/2022

1 exp Premature Birth/ or exp Infant, Premature/ or Infant, Newborn

2 exp kidney/

3 structure.mp

4 morphology.mp

5 3 or 4

6 exp biopsy

7 exp autopsy

8 exp histology

9 exp immunohistochemistry

10 6 or 7 or 8 or 9

11 exp magnetic resonance imaging

12 exp X‐rays

13 exp ultrasonography

14 exp ultrasonography, Doppler, Color/ or exp ultrasonography, Doppler/ or exp ultrasonography, Dopper, Duplex

15 exp nuclear medicine

16 exp Radionuclide imaging

17 exp Tomography, X‐ray computed

18 exp Fluoroscopy

19 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18

20 10 or 19

21 renal.mp

22 2 or 21

23 1 and 5 and 20 and 22

Google scholars

Last search 21/09/2022

Kidney, renal, structure, morph*, vasculature, scintigraphy, biopsy, imaging, ultrasound, xray, mri, newborn, neonates

Search results were sorted by relevance and the pages screened by title search until >10 pages no hits.

CINAHL

Latest search 31/08/2022

Neonate or neonatal or premature or preterm or newborn

AND kidney OR renal

AND structure OR morphology

AND ultrasound or sonography or sonogram or ultrasonography OR duplex doppler OR MRI or magnetic resonance imaging or mri scan OR computed tomography or ct or cat scan OR Xray or x‐ray or xrays or x‐rays or radiograph OR scintigraphy OR nuclear medicine imaging OR biopsy

Staub, E. (2025). Current and potential methods to assess kidney structure and morphology in term and preterm neonates. The Anatomical Record, 308(4), 1229–1250. 10.1002/ar.25195