Quantitative MRI of fatty liver disease in a large pediatric cohort: correlation between liver fat fraction, stiffness, volume, and patient-specific factors

Madalsa Joshi; Jonathan R. Dillman; Kamalpreet Singh; Suraj D. Serai; Alexander J. Towbin; Stavra Xanthakos; Bin Zhang; Weizhe Su; Andrew T. Trout

doi:10.1007/s00261-017-1289-y

Abdominal Radiology

May 2018, Volume 43, Issue 5, pp 1168–1179 | Cite as

Quantitative MRI of fatty liver disease in a large pediatric cohort: correlation between liver fat fraction, stiffness, volume, and patient-specific factors

Authors
Authors and affiliations

Madalsa Joshi
Jonathan R. DillmanEmail author
Kamalpreet Singh
Suraj D. Serai
Alexander J. Towbin
Stavra Xanthakos
Bin Zhang
Weizhe Su
Andrew T. Trout

Article

First Online: 21 August 2017

5 Shares
108 Downloads
1 Citations

Abstract

Purpose

Magnetic resonance imaging (MRI) techniques are increasingly used to quantify and monitor liver tissue characteristics including fat fraction, stiffness, and liver volume. The purpose of this study was to assess the inter-relationships between multiple quantitative liver metrics and patient-specific factors in a large pediatric cohort with known or suspected fatty liver disease.

Materials and methods

In this IRB-approved, HIPAA-compliant study, we retrospectively reviewed patient data and quantitative liver MRI results in children with known/suspected fatty liver disease. Relationships between liver MRI tissue characteristics and patient variables [sex, age, body mass index (BMI), diabetic status (no diabetes mellitus, insulin resistance/“prediabetes” diagnosis, or confirmed diabetes mellitus), and serum alanine transaminase (ALT)] were assessed using linear mixed models.

Results

294 quantitative liver MRI examinations were performed in 202 patients [128/202 (63.4%) boys], with a mean age of 13.4 ± 2.9 years. Based on linear mixed models, liver fat fraction was influenced by age (−0.71%/+1 year, p = 0.0002), liver volume (+0.006%/+1 mL, p < 0.0001), liver stiffness (−2.80%/+1 kPa, p = 0.0006), and serum ALT (+0.02%/+1 U/L, p = 0.0019). Liver stiffness was influenced by liver volume (+0.0003 kPa/+1 mL, p = 0.001), fat fraction (−0.02 kPa/+1% fat, p = 0.0006), and ALT (0.002 kPa/+1 U/L, p = 0.0002). Liver volume was influenced by sex (−262.1 mL for girls, p = 0.0003), age (+51.8 mL/+1 year, p = 0.0001), BMI (+49.1 mL/+1 kg/m², p < 0.0001), fat fraction (+30.5 mL/+1% fat, p < 0.0001), stiffness (+192.6 mL/+1 kPa, p = 0.001), and diabetic status (+518.94 mL for diabetics, p = 0.0009).

Conclusions

Liver volume, fat fraction, and stiffness are inter-related and associated with multiple patient-specific factors. These relationships warrant further study as MRI is increasingly used as a non-invasive biomarker for fatty liver disease diagnosis and monitoring.

Keywords

MRI Children Steatosis Fibrosis Quantitative Multiparameteric

This is a preview of subscription content, to check access.

Notes

Compliance with ethical standards

Funding

None.

Conflict of interest

Joshi—none, Dillman—unrelated grant funding (Siemens Medical Solutions USA, Toshiba America Medical Systems, Guerbet Group), Singh—none, Serai—None, Towbin—unrelated grant funding (Siemens Medical Solutions USA, Guerbet Group), consultant (Applied Radiology), and royalties (Elsevier), Xanthakos—none, Zhang—none, Shu—none, Trout—unrelated grant funding (Siemens Medical Solutions USA, Toshiba America Medical Systems), consultant (American College of Radiology), and royalties (Elsevier).

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. As this study was retrospective and involved analysis of existing data, the requirement for informed consent was waived by our institutional review board.

Appendix 1

MR sequence parameters for fat fraction sequence on Philips and GE 1.5T MRI scanners.

Scanner	Philips Ingenia	Philips Ingenia	GE HDx
Technique	mDixon quant	Chemical shift encoded with lipid bag	Chemical shift encoded with lipid bag
Pulse sequence	3D GRE	3D GRE	3D GRE
Matrix	160 × 140	256 × 192	256 × 192
# of averages	1	1	1
First TE (msec)	0.98	2.2	2.2
Delta TE (msec)	0.7	2.2	2.2
# of echoes	6	2	2
TR (msec)	5.7	6.7	6.7
Flip angle (degrees)	5	5	5
Bandwidth (Hz/px)	2799	500	500
Slice thickness (mm)	6	4	4
Acceleration	SENSE	NA	NA
Acceleration factor	2	NA	NA
Scan time (minutes)	0:16	0:18	0:18

GRE gradient recalled echo; TE echo time; TR repetition time; Hz hertz; px pixel

Appendix 2

Open image in new window

An example of chemical shift fat fraction calculation with correction to a 20% lipid standard is shown. Opposed-phase (A1) and in-phase (A2) images of the liver with the lipid bag (Fresenius Kabi, Sweden) along the left abdominal wall are also shown. Regions of interest (ROIs) of approximately 200–300 mm² are shown, placed both in the right hepatic lobe, avoiding large vessels (black circles), and in the lipid bag on the same images (white circles). Corrected chemical shift is calculated as follows (LB = lipid bag, IP = in phase, OP = opposed phase):

$$ {\text{Correction factor }} = 0.2/(\frac{LB IP - LB OP}{{2\left( {LB IP} \right)}}) $$

$$ {\text{Uncorrected liver fat }} = \frac{Liver IP - Liver OP}{{2\left( {Liver IP} \right)}} $$

$$ {\text{Corrected liver fat }} = {\text{ Uncorrected liver fat }} \times {\text{ Correction factor}}. $$

Corrected liver fat is measured on each of three images and reported as an average of those three values. Uncorrected and corrected fat fraction values measured using this technique were compared to the results obtained with a commercially available gradient-echo-based fat quantitation sequence (HepaFat-Scan, Resonance Health; Australia) in 11 patients (A3). There was a strong, statistically significant correlation between each of the chemical shift techniques and the commercially available sequence (r = 0.96, p = 0.006 for both comparisons). The uncorrected data are indicated by empty boxes with a thin dashed trend line. The corrected data are indicated by black diamonds with a thin solid trend line. The solid gray line is the line of equality. For the exam shown, chemical shift fat fraction was calculated at 10.2% with a fat fraction of 9.6% calculated by HepaFat.

Appendix 3

Linear mixed model parameter estimates and p values based on solution for fixed effects for prediction of liver fat fraction (%). Models include variable (Age*Age_group), where 1 = age >10 years. Data for two separate models (“Clinical Data Only” and “Clinical and Imaging Data”) are presented.

Parameter	Clinical data only		Clinical and imaging data
Parameter	Estimate	p value	Estimate	p value
Sex (female)	−3.44	0.0013	−1.16	0.2586
Age (years)	0.01	0.9961	−0.45	0.2281
Age*Age_group	−0.30	0.1851	−0.16	0.4299
BMI (kg/m²)	0.19	0.0189	−0.10	0.2827
ALT (U/L)	0.03	0.0001	0.02	0.0016
Diabetes	3.66	0.1211	0.17	0.9389
“Prediabetes”	0.43	0.6810	0.35	0.7164
Liver volume (mL)			0.006	<.0001
Liver stiffness (kPa)			−2.70	0.0012

Linear mixed model parameter estimates and p values based on solution for fixed effects for prediction of liver stiffness (kPa). Models include variable (Age*Age_group), where 1 = age >10 years. Data for two separate models (“Clinical Data Only” and “Clinical and Imaging Data”) are presented.

Parameter	Clinical data only		Clinical and imaging data
Parameter	Estimate	p value	Estimate	p value
Sex (female)	0.04	0.5770	0.10	0.2499
Age (years)	−0.04	0.1295	−0.08	0.0111
Age*Age_group	0.03	0.0420	0.04	0.0390
BMI (kg/m²)	0.02	<.0001	0.01	0.1685
ALT (U/L)	0.002	<.0001	0.002	0.0006
Diabetes	0.21	0.1633	0.12	0.5119
“Prediabetes”	0.04	0.5322	0.06	0.4728
Liver volume (mL)			0.0003	0.0012
Fat fraction (%)			−0.02	0.0012

Linear mixed model parameter estimates and p values based on solution for fixed effects for prediction of liver volume (mL). Models include variable (Age*Age_group), where 1 = age >10 years. Data for two separate models (“Clinical Data Only” and “Clinical and Imaging Data”) are presented.

Parameter	Clinical data only		Clinical and imaging data
Parameter	Estimate	p value	Estimate	p value
Sex (female)	−288.19	<.0001	−262.08	0.0003
Age (years)	13.42	0.5971	52.78	0.0487
Age*Age_group	12.71	0.3586	−0.65	0.9652
BMI (kg/m²)	58.57	<.0001	49.08	<.0001
ALT (U/L)	2.21	<.0001	0.74	0.1531
Diabetes	560.05	0.0002	519.21	0.0009
“Prediabetes”	46.22	0.4939	−2.35	0.9731
Fat fraction (%)			30.50	<.0001
Mean stiffness (kPa)			192.98	0.0012

References

1.
Kramer H, Pickhardt PJ, Kliewer MA, et al. (2017) Accuracy of liver fat quantification with advanced CT, MRI, and ultrasound techniques: prospective comparison with MR spectroscopy. AJR Am J Roentgenol. 208(1):92–100CrossRefPubMedGoogle Scholar
2.
Petitclerc L, Sebastiani G, Gilbert G, et al. (2017) Liver fibrosis: review of current imaging and MRI quantification techniques. J Magn Reson Imaging. 45(5):1276–1295CrossRefPubMedGoogle Scholar
3.
Srinivasa Babu A, Wells ML, Teytelboym OM, et al. (2016) Elastography in chronic liver disease: modalities, techniques, limitations, and future directions. Radiographics. 36(7):1987–2006CrossRefPubMedPubMedCentralGoogle Scholar
4.
Yin M, Glaser KJ, Talwalkar JA, et al. (2016) Hepatic MR elastography: clinical performance in a series of 1377 consecutive examinations. Radiology. 278(1):114–124CrossRefPubMedGoogle Scholar
5.
Towbin AJ, Serai SD, Podberesky DJ (2013) Magnetic resonance imaging of the pediatric liver: imaging of steatosis, iron deposition, and fibrosis. Magn Reson Imaging Clin N Am. 21(4):669–680CrossRefPubMedGoogle Scholar
6.
Trout AT, Serai S, Mahley AD, et al. (2016) Liver stiffness Measurements with MR elastography: agreement and repeatability across imaging systems, field strengths, and pulse sequences. Radiology. 281(3):793–804CrossRefPubMedGoogle Scholar
7.
Manduca A, Oliphant TE, Dresner MA, et al. (2001) Magnetic resonance elastography: non-invasive mapping of tissue elasticity. Med Image Anal. 5(4):237–254CrossRefPubMedGoogle Scholar
8.
Pavlides M, Banerjee R, Tunnicliffe EM, et al. (2016) Multi-parametric magnetic resonance imaging for the assessment of non-alcoholic fatty liver disease severity. Liver Int. 37(7):1065–1073CrossRefGoogle Scholar
9.
Leitao HS, Doblas S, Garteiser P, et al. (2017) Hepatic fibrosis, inflammation, and steatosis: influence on the MR viscoelastic and diffusion parameters in patients with chronic liver disease. Radiology. 283(1):98–107CrossRefPubMedGoogle Scholar
10.
Loomba R, Wolfson T, Ang B, et al. (2014) Magnetic resonance elastography predicts advanced fibrosis in patients with nonalcoholic fatty liver disease: a prospective study. Hepatology. 60(6):1920–1928CrossRefPubMedPubMedCentralGoogle Scholar
11.
Chen J, Talwalkar JA, Yin M, et al. (2011) Early detection of nonalcoholic steatohepatitis in patients with nonalcoholic fatty liver disease by using MR elastography. Radiology. 259(3):749–756CrossRefPubMedPubMedCentralGoogle Scholar
12.
Patel NS, Hooker J, Gonzalez M, et al. (2017) Weight loss decreases magnetic resonance elastography estimated liver stiffness in nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol. 15(3):463–464CrossRefPubMedGoogle Scholar
13.
Patel NS, Doycheva I, Peterson MR, et al. (2015) Effect of weight loss on magnetic resonance imaging estimation of liver fat and volume in patients with nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol. 13(3):561–568CrossRefPubMedGoogle Scholar
14.
Macaluso FS, Maida M, Camma C, et al. (2014) Steatosis affects the performance of liver stiffness measurement for fibrosis assessment in patients with genotype 1 chronic hepatitis C. J Hepatol. 61(3):523–529CrossRefPubMedGoogle Scholar
15.
Petta S, Maida M, Macaluso FS, et al. (2015) The severity of steatosis influences liver stiffness measurement in patients with nonalcoholic fatty liver disease. Hepatology. 62(4):1101–1110CrossRefPubMedGoogle Scholar
16.
Conti F, Vukotic R, Foschi FG, et al. (2016) Transient elastography in healthy subjects and factors influencing liver stiffness in non-alcoholic fatty liver disease: an Italian community-based population study. Dig Liver Dis. 48(11):1357–1363CrossRefPubMedGoogle Scholar
17.
Guo Y, Dong C, Lin H, et al. (2017) Ex vivo study of acoustic radiation force impulse imaging elastography for evaluation of rat liver with steatosis. Ultrasonics. 74:161–166CrossRefPubMedGoogle Scholar
18.
Kang BK, Lee SS, Cheong H, et al. (2015) Shear wave elastography for assessment of steatohepatitis and hepatic fibrosis in rat models of non-alcoholic fatty liver disease. Ultrasound Med Biol. 41(12):3205–3215CrossRefPubMedGoogle Scholar
19.
Silva M, Marques M, Cardoso H, et al. (2016) Glycogenic hepatopathy in young adults: a case series. Rev Esp Enferm Dig. 108(10):673–676PubMedGoogle Scholar
20.
DiPaola FW, Schumacher KR, Goldberg CS, et al. (2017) Effect of Fontan operation on liver stiffness in children with single ventricle physiology. Eur Radiol. 27(6):2434–2442CrossRefPubMedGoogle Scholar
21.
Serai SD, Dillman JR, Trout AT (2017) Proton density fat fraction measurements at 1.5- and 3-T hepatic MR imaging: same-day agreement among readers and across two imager manufacturers. Radiology. 284(1):244–254. doi: 10.1148/radiol.2017161786 CrossRefPubMedGoogle Scholar

Copyright information

Authors and Affiliations

Madalsa Joshi
- 1
Jonathan R. Dillman
- 1
Email author
Kamalpreet Singh
- 2
Suraj D. Serai
- 1
Alexander J. Towbin
- 1
Stavra Xanthakos
- 3
Bin Zhang
- 4
Weizhe Su
- 5
Andrew T. Trout
- 1

1.Department of RadiologyCincinnati Children’s Hospital Medical CenterCincinnatiUSA
2.Department of RadiologyAllegheny General HospitalPittsburghUSA
3.Division of Gastroenterology, Hepatology and Nutrition, Department of PediatricsCincinnati Children’s Hospital Medical CenterCincinnatiUSA
4.Division of Biostatistics and EpidemiologyCincinnati Children’s Hospital Medical CenterCincinnatiUSA
5.Department of Mathematical SciencesUniversity of CincinnatiCincinnatiUSA

Quantitative MRI of fatty liver disease in a large pediatric cohort: correlation between liver fat fraction, stiffness, volume, and patient-specific factors

Abstract

Purpose

Materials and methods

Results

Conclusions

Keywords

Notes

Compliance with ethical standards

Funding

Conflict of interest

Ethical approval

References

Copyright information

Authors and Affiliations

Personalised recommendations

Cite article

Cookies