ORIGINAL ARTICLE

Pediatric Liver MR Elastography

Suraj D. Serai • Alexander J. Towbin •

Daniel J. Podberesky

Received: 19 December 2011 / Accepted: 14 April 2012 / Published online: 9 May 2012

� Springer Science+Business Media, LLC 2012

Abstract

Introduction Many chronic pediatric liver disorders are

complicated by the development of fibrosis and ultimately

cirrhosis. Although hepatic fibrogenesis progresses along a

common pathway irrespective of the specific etiology,

fibrosis in pediatric liver diseases has different histopathol-

ogical patterns than in adults. In pediatric liver disease, as in

adults, management choices may depend upon the stage of

fibrosis at diagnosis. With early intervention, the progression

of hepatic fibrosis can be slowed or halted, and in some

situations, reversed. While liver biopsy is the gold standard

for diagnosing and assessing the presence and degree of

fibrosis, it has several disadvantages including the potential

for sampling error, the risk of complications, the relatively

high cost, and general poor acceptance by pediatric patients

and their parents. MR elastography (MRE) is a relatively

new imaging technique with the potential for allowing a

safe, rapid, cost-effective, and non-invasive evaluation of a

wide variety of hepatic diseases by quantitatively evaluating

the stiffness of the liver parenchyma. The purpose of this

article is to present our initial clinical experience and illus-

trate our modified technique for the application of liver

MRE in pediatric patients at our medical center.

Methods and Materials Pediatric MRE techniques were

developed and applied to over 45 patients scanned with our

new protocol.

Conclusion Liver MRE is a safe, non-invasive method

for assessing hepatic fibrosis in pediatric patients.

Keywords Pediatric � Elastography � MRE � Fibrosis �Steatohepatitis

Introduction

The most common causes of chronic liver diseases in children

are hepatitis (infectious, autoimmune, drug-related), genetic

diseases (a-1 antitrypsin deficiency, cystic fibrosis, biliary

atresia, Wilson disease, storage disorders), non-alcoholic fatty

liver disease (NAFLD) and non-alcoholic steatohepatitis

(NASH) [1]. While there are many potential etiologies for

pediatric liver disease, if untreated, they can all progress to

hepatic fibrosis, and eventually cirrhosis [2, 3]. If the under-

lying cause of liver disease can be effectively treated, there is a

high probability that with early intervention, the degree of

fibrosis can be minimized and possibly reversed [4–6]. Hence,

early diagnosis, follow-up, and therapeutic monitoring of

fibrogenesis, the process of generation of new connective

tissue in a diseased liver, is of great clinical importance.

Traditionally, liver biopsy has been used as the gold

standard for staging and grading hepatic fibrosis. While

biopsy has been a useful method for diagnosing and staging

liver disease for many patients, there are several limitations

of this procedure, including its invasive nature, potential for

sampling error, relatively high cost, and risk of life-threat-

ening complications such as hemorrhage, infection, and

organ damage. Liver biopsies are also subject to inter- and

intra-observer variability [7, 8]. These limitations highlight

the need for a non-invasive test that is able to accurately and

objectively measure the degree of hepatic fibrosis, especially

in children, for whom the risk/benefit ratio of biopsy must be

even more carefully weighed than in adults [9].

A number of non-invasive imaging techniques have been

developed in an attempt to diagnose hepatic fibrosis,

S. D. Serai (&) � A. J. Towbin � D. J. Podberesky

Department of Radiology, MLC 5031, Cincinnati Children’s

Hospital and Medical Center, 3333 Burnet Ave., Cincinnati,

OH 45229, USA

e-mail: suraj.serai@cchmc.org

123

Dig Dis Sci (2012) 57:2713–2719

DOI 10.1007/s10620-012-2196-2

including diffusion MR, magnetization transfer technique,

non-contrast enhanced CT, and transient ultrasound elas-

tography [10–13]. MR elastography (MRE) of the liver has

recently emerged as a novel technique for providing a

quantitative assessment of liver stiffness. MRE assesses

tissue stiffness by measuring the speed of shear waves

propagating within the parenchyma. This technique is FDA

approved, and in adults has been shown to accurately detect

and stage hepatic fibrosis and accurately detect steatohep-

atitis [14–16]. However, there is very limited published

information on pediatric applications of liver MRE [17].

The purpose of this article is to present our initial clinical

experience, a detailed description of the application of liver

MRE, and illustrate our modified technique for the appli-

cation of liver MRE in pediatric patients.

Methods

MRE equipment consists of an active driver kept in the MR

equipment room and a passive driver that is placed on the

patient during imaging in the MRI scanner (Fig. 1). The

passive driver is approximately 7 in. in diameter and is

connected to the active driver in the equipment room by a wave-

guide with a hollow plastic tube. The audio subwoofer magnets,

which act as the active driver, generate 60-Hz vibrations, which

are passed via pneumatic pressure to the passive driver. The

initiation and cessation of the vibrations are controlled by the

MR pulse sequence. Our MRE protocol was originally adapted

and modified from adult MRE scan methods developed by

Ehman et al. at the Mayo Clinic [15–19].

In order to prepare pediatric patients for the sensation of

the hepatic vibrations, child life specialists use a vibrating

passive driver simulator on the child prior to the MRE.

This pre-scan simulation helps prepare patients for the

sensation they will feel during scanning and thus reduces

anxiety and sudden movements at the start of the actual

MRE sequence. Patient well-being is monitored throughout

the scan with intercom communication between all imag-

ing series. In addition, the patient is instructed to use the

squeeze-ball alarm at the onset of any heating, discomfort,

or any other unusual sensation that occurs during scanning.

Fig. 1 a A schematic diagram

of patient setup with the MRE

hardware. The active driver is

placed in the MR computer

room and the passive driver,

connected via a plastic tube

through a wave-guide, is

positioned on the anterior body

approximately over the liver

region (adapted and modified

from Yin et al. [17]).

b Photograph of a patient being

positioned for liver MRE exam

2714 Dig Dis Sci (2012) 57:2713–2719

123

Liver MRE can be performed on sedated or non-sedated

children. While in adults it is recommended that liver MRE

be performed after an overnight fast, in our protocol for

children we ask that they take nothing by mouth for at least

4 h before the MRE exam [18, 19].

MRE pulse sequence techniques originally derived from

adult applications were adapted and modified to fit pediatric

needs. As with other routine pediatric body protocols, the

specific absorption rate (SAR) is automatically maintained

within acceptable limits determined by the MR scanner based

upon the entered patient weight. In addition to standard

adjustments in scan field-of-view (FOV), the driver power

level is reduced by 20 % in our typical patients (ages

5–18 years) and 40–50 % in very young children (\2 years

old) in order to prevent any theoretical injury due to the

vibrating driver. For patients between ages 2 and 5 years, an

appropriate intermediate power level can be subjectively

selected based on patient height, weight, and size.

In order to maintain image quality and further minimize

any potential risk of mechanical injury to the child due to the

vibrations, a folded towel is placed between the passive

driver and the child’s abdomen (Fig. 1b). This helps to

reduce the air space void and increases the mechanical

coupling between the thoraco-abdominal wall and the

comparatively large passive driver. For very young patients,

the modified MRE pulse sequence triggered vibrations have

not been limiting even though only a portion of the

abdominal wall is in contact with the passive driver.

In our practice, MRE is typically performed as part of a

diagnostic liver MRI exam. However, we have also created a

fast, relatively low-cost limited liver MRE examination. The

different clinical imaging protocols of routine liver imaging

with MRE exam and the fast limited MRE exam are descri-

bed in detail in Table 1. To date, at our imaging center, liver

MRE has been performed in over 45 pediatric patients (age

range 6–17 years). All studies were performed on a 1.5-T GE

HDXt MRI scanner (General Electric, Waukesha, WI). For

the MRE sequence, four axial slices through the liver, each

8–10 mm thick, are prescribed from the coronal T1 W

localizer sequence. When prescribing the axial slices, the

technologist chooses a location inferior to the heart at the

widest portion of the liver (Fig. 2). Care is taken to ensure

that the localizer is obtained at end expiration so that the

prescribed slices are placed in a reproducible location.

Image acquisition for the MRE pulse sequence is com-

pleted in four breath-holds (each lasting approximately

12–15 s) performed at end expiration. If the patient cannot

hold his or her breath, the scan may be performed while

free breathing with a single average rather than multiple

averages. In our experience, increasing the number of

averages may result in an incorrect masking when the

elastograms are overlaid with the magnitude image for

stiffness measurements (Fig. 3). For each slice scanned,

four magnitude and four phase images are obtained; thus a

four-slice acquisition results in 16 magnitude images and

16 phase images.

Table 1 Clinical protocols for liver MRE procedure

Sequence Plane Approx.

scan time (s)

TR (ms) TE

(ms)

Matrix

size

Slice thickness

(mm)

GAP

(mm)

Protocol routine

liver with MRE

Routine liver with MRE

2D T2WFS Axial 160 3,000 85 256 9 192 5 1 T2W Fat Sat

2D T1W Axial 185 400 Min 256 9 192 5 1

MRE FGRE Axial 50 50 Min 256 9 64 10 1 Elastography sequence

Balanced SSFP FS Axial 90 Min 224 9 224 5 1 Balanced weighting—axial

Balanced SSFP FS Coronal 90 Min 256 9 256 4 1 Balanced weighting—coronal

T2W Coronal 180 3,000 85 256 9 192 5 1 T2W

2D DWI Axial 80 3,000 Min 192 9 128 5 1 Diffusion-weighted imaging

2D TOF venous 230 34 18 256 9 128 2 0 TOF venography

Post-contrast MRA 120 4 Min 256 9 192 2.5 0 TOF angiography

T1W Post-contrast FS Axial 185 400 Min 256 9 192 5 1 T1W Fat Sat—post-contrast

3D GRE (In/out phase) Axial 21 180 Min 320 9 192 3 0 T1W in-phase and opposed phase

Limited liver MRE

MRE FGRE Axial 50 50 Min 256 9 64 10 1 Elastography sequence

3D GRE (in/out phase) Axial 21 180 Min 320 9 192 3 0 T1W in-phase and opposed phase

2D T1W Coronal 185 400 Min 256 9 192 5 1 T1W

2D T2WFS Axial 150 3,000 85 256 9 192 5 1 T2W Fat Sat

MRE FGRE MR elastography fast-gradient echo sequence, FS fat saturation included, TOF MRV time of flight MR venography, MRA MR

angiography, Min minimum, SSFP steady-state free precession

Dig Dis Sci (2012) 57:2713–2719 2715

123

We have found that there are several potential causes

of error when performing the MRE sequences, such as

disconnection of the pneumatic tubing, incorrect place-

ment of the passive driver, power supply of the active

driver being turned off, and incorrect pulse sequence

download. Therefore, we check the phase images once

the sequence is complete to ensure correct image acqui-

sition. Examples of a typical phase image and a phase

image obtained after incorrect pulse sequence download

are shown in Fig. 4.

Fig. 2 MRE slice selection:

The axial slices are prescribed

(red lines) so that the liver is

imaged in its widest portion,

inferior to the heart

Fig. 3 MRE stiffness map

(scale 0–8 kPa) of a

volunteer comparing a

breath-hold (BH) versus

free-breathing (FB) with

number of averages (NX) 1,

2, and 4, respectively. More

averages with free-breathing

can result in incorrect

masking of elastograms

during the ROI selection

with the magnitude images

2716 Dig Dis Sci (2012) 57:2713–2719

123

After the MRE magnitude and phase images are

obtained, the data is automatically converted at the console

into images displaying the stiffness of the liver paren-

chyma. Stiffness maps, referred to as elastograms, and

quantitative mean stiffness values are generated from the

phase images. A basic qualitative elastogram can be gen-

erated by the technologist on the scanner console. In order

to obtain a quantitative elastogram with stiffness values,

the phase images are sent to another computer for post-

processing (MRE Wave; Mayo Clinic, Rochester, NY).

The first step of post-processing is to draw a region-of-

interest (ROI) around the liver on each of the four ana-

tomical image slices (Fig. 5). Care must be taken while

drawing this ROI to stay within the liver parenchyma and

to carefully avoid large blood vessels. The anatomic ROI is

then combined with the wave images to identify areas

where wave penetration is not observed. If needed, the ROI

can be modified to exclude these areas. Once the final ROI

is identified, the software generates the quantitative stiff-

ness values. These stiffness values, along with published

liver reference values, can be used to determine the pres-

ence and degree of liver fibrosis or steatohepatitis [18, 19].

Discussion

In our experience thus far, MRE has been helpful in

assessing the liver in pediatric patients with elevated liver

enzymes from unknown etiology, patients with suspected

hepatic masses, suspected hepatitis, NAFLD, inflammatory

bowel disease, partial hepatic resections, congenital fibro-

genic liver diseases, worsening ascites, biliary atresia,

storage disorders, cardiac failure and cystic fibrosis

(Fig. 6). MRE has helped us prevent several liver biopsies

Fig. 4 Example of MRE phase image error. The waves transmitted

by the passive driver are seen from the anterior end of the liver

(arrow). a These waves are expected to be seen in the phase images.

b No waves are seen in this phase image. This error occurred because

of a communication failure between the pulse sequence and the

passive driver

Fig. 5 Example post-processing

analysis of ROI determination on

elastogram and magnitude

image. The ROI (arrow) is

carefully chosen to include liver

only. The stiffness value is

computed from regions within

the ROI

Dig Dis Sci (2012) 57:2713–2719 2717

123

in NAFLD patients at our hospital after normal results were

obtained from the elastograms (Fig. 7).

Conclusions

Liver MRE is a safe, non-invasive method for assessing

hepatic fibrosis. With our modified technique, we have

been able to successfully perform liver MRE in children.

Further investigation is planned in order to obtain normal

liver stiffness values in children, and to determine the

sensitivity of liver MRE for identifying and staging hepatic

fibrosis in children.

Acknowledgments The authors would like to thank Drs. Richard L.

Ehman and Meng Yin of the Mayo Clinic for their assistance and

support in establishing our pediatric liver MRE program.

Conflict of interest None.

References

1. Liver disorders in childhood 3rd edition. Oxford, Butterworth-

Heinemann Ltd, 1994, pp. 180–196.

2. Mieli-Vergani G, Vergani D. Autoimmune pediatric liver disease.

World J Gastroenterol. 2008;14:3360–3367.

3. Brunt EM. Pathology of nonalcoholic fatty liver disease. Nat RevGastroenterol Hepatol. 2010;7:195–203.

Fig. 6 A 12-year-old male with

a history of tetralogy of Fallot

with pulmonary atresia,

hypoplastic right ventricle:

a elastogram and b magnitude

images show an unexpected

elevation in liver stiffness

(calculated to be 3.5 kPa)

presumed to be due to fibrosis

from passive congestion. The

artifact (arrow) seen in the

magnitude image is due to

sternotomy wires

Fig. 7 A 17-year-old male with

a history of NAFLD and

elevated liver enzymes.

a Inphase and b Out-of-phase.

Opposed phase images show

signal drop-out consistent with

diffuse fatty infiltration of the

liver. Normal MRE stiffness

value of 2.31 kPa was obtained

from the c magnitude and

d elastogram images, avoiding

the need for a liver biopsy

2718 Dig Dis Sci (2012) 57:2713–2719

123

4. Friedman SL. Hepatic fibrosis—overview. Toxicology. 2008;254:

120–129.

5. Friedman SL. Evolving challenges in hepatic fibrosis. Nat RevGastroenterol Hepatol. 2010;7:425–436.

6. Bortolotti F, Guido M. Reversal of liver cirrhosis: a desirable

clinical outcome and its pathogenic background. J Pediatr Gas-troenterol Nutr. 2007;44:401–406.

7. Pariente D, Franchi-Abella S. Pediatric chronic liver diseases:

how to investigate and follow up? Role of imaging in the diag-

nosis of fibrosis. Pediatr Radiol. 2010;40:906–919.

8. Castera L, Pinzani M. Biopsy and non-invasive methods for the

diagnosis of liver fibrosis: does it take two to tango? Gut.2010;59:861–866.

9. Straub BK, Schirmacher P. Pathology and biopsy assessment of

non-alcoholic fatty liver disease. Dig Dis. 2010;28:197–202.

10. Castera L. Non-invasive diagnosis of steatosis and fibrosis.

Diabetes Metab. 2008;34:674–679.

11. Chen JH, Chai JW, Shen WC. Magnetization transfer contrast

imaging of liver cirrhosis. Hepatogastroenterology. 1999;46:

2872–2877.

12. Aube C, Racineux PX, Lebigot J, et al. Diagnosis and quantifi-

cation of hepatic fibrosis with diffusion weighted MR imaging:

preliminary results. J Radiol. 2004;85:301–306.

13. Yeh WC, Li PC, Jeng YM, et al. Elastic modulus measurements

of human liver and correlation with pathology. Ultrasound MedBiol. 2002;28:467–474.

14. Shannon A, Alkhouri N, et al. Ultrasonographic quantitative

estimation of hepatic steatosis in children with NAFLD. J PediatrGastroenterol Nutr. 2011;53(2):190–195.

15. Muthupillai R, Rossman PJ, et al. Magnetic resonance elastog-

raphy by direct visualization of propagating acoustic strain

waves. Science. 1995;269:1854–1857.

16. Mariappan YK, Glaser KL, Ehman RL. Magnetic resonance

elastography: a review. Clin Anatomy. 2010;23:497–511.

17. Binkovitz LA, El-Youssef M, Gaser KJ, Yin M, Binkovitz AK,

Ehman RL. Pediatric MR elastography of hepatic fibrosis: prin-

ciples, technique and early clinical experience. Pediatr Radiol.2012;42(4):402–409.

18. Yin M, Talwalkar JA, et al. Assessment of hepatic fibrosis with

magnetic resonance elastography. Clin Gastroenterol Hepatol.2007;5:1207–1213.

19. Yin M, Talwalkar JA, et al. Dynamic postprandial hepatic stiff-

ness augmentation assessed with MR elastography in patients

with chronic liver disease. AJR. 2011;197:64–70.

Dig Dis Sci (2012) 57:2713–2719 2719

123