Quantification of pancreatic fat with dual-echo imaging at 3.0-T MR in clinical application: how do the corrections for T1 and T2* relaxation effect work and simplified correction strategy

Abstract

Background

Dual-echo imaging is a routine clinical magnetic resonance (MR) sequence affected by T1 and T2* relaxation effect in fat quantification. The separate impacts of T1 and T2* relaxation effect in pancreatic fat quantification using dual-echo imaging at 3.0-T MR have not been reported in detail.

Purpose

To demonstrate the separate T1 and T2* relaxation effect on pancreatic fat quantification by dual-echo imaging at 3.0-T MR and the simplified correction strategy is discussed for convenient clinical application.

Material and Methods

Twenty-one non-alcoholic fatty liver disease (NAFLD) participants with high risk of pancreatic steatosis were included. Pancreatic fat fractions (FF) by dual-echo imaging with different corrections were compared to that of proton magnetic resonance spectroscopy (¹H-MRS). Correlation analysis and Bland–Altman analysis were applied.

Results

The FF by ¹H-MRS was 5.9 ± 1.7%. Significant positive correlation (all P < 0.01) was found between FF by ¹H-MRS and each dual-echo imaging, in which T1 and T2* correction showed the best correlation (r = 0.95, FF = 6.2 ± 1.7%) and no correction showed the worst correlation (r = 0.86, FF = 5.2 ± 2.0%), and the simplified T1 and T2* correction manifested as r = 0.93 and FF = 6.3 ± 1.8%. FF by T1 and T2* correction showed the best agreement, while T1 correction showed the worst agreement as compared to that of ¹H-MRS.

Conclusion

T1 and T2* correction shows the best performance while no correction dual-echo imaging remains clinical available which may benefit from prior OP echo. Simplified correction using single T2* (32.6 ms) of water and fat is recommended for convenient clinical application in absence of obvious pancreatic iron overload.

Keywords

Abdomen/GI magnetic resonance imaging (MRI)pancreas adults technology assessments

Introduction

Pancreatic steatosis can be induced by numerous conditions including obesity, diabetes, dietary deficiency, and hormonal therapy (1 –3). It is considered an analog of non-alcoholic fatty liver disease (NAFLD) and the pancreatic manifestation of “metabolic syndrome” (4). Pancreatic steatosis contributes to decreased β-cell mass and function, possibly by local release of non-esterified fatty acids and adipocyte-derived proinflammatory and vasoactive factors. Collectively, these may impair β-cells, thereby causing dysfunction (5). Therefore, it is important to find a convenient and accurate method to monitor pancreatic fat in people with high-risk factors.

Biopsy is the traditional gold standard for fat content measurement but is impeded by its invasiveness and sampling error (6). Magnetic resonance imaging (MRI) methods such as chemical shift imaging and proton magnetic resonance spectroscopy (¹H-MRS) have emerged as non-invasive approaches for in vivo fat detection and quantification. ¹H-MRS is highly accurate and generally considered the gold standard for fat quantification in many organs (7 –10). ¹H-MRS yields direct information of chemical species with high spectral resolution, but is usually performed in one voxel at a time and does not provide any anatomical image information. In addition, ¹H-MRS is time-consuming for routine clinical applications in the abdomen and requires highly skilled operators for post-processing. Moreover, standard ¹H-MRS has relatively long minimum echo times (about 20 and 30 ms for STEAM and PRESS, respectively), which may introduce T2 weighting between the signals of individual peaks (11). Based on the chemical-shift imaging, the dual-echo imaging (Dixon) is practical for fat detection and quantification by using in-phase (IP) and opposed-phase (OP) images (12). Dual-echo imaging provides accurate anatomical images and covers the entire imaging field allowing fat quantification in simple post-processing. Scan time for dual-echo imaging is much shorter than ¹H-MRS. However, dual-echo imaging could be affected by T1 and T2* relaxation effects leading to lower accuracy of fat quantification (13).

Currently, many methods can compensate for T1 and T2* relaxation effects in dual-echo imaging, such as low flip angle (FA) or dual FA to minimize the T1 bias (14,15), multi-echo to measure T2* of water and fat separately (16,17) or the equal T2* of water and fat (15,18) for correcting T2* relaxation effect. However, these methods are complicated and rarely applied in clinical practice. The integrated MR sequences like Multi-Echo 3D GRE (Siemens) and IDEAL-IQ (GE) can correct the effects and give accurate fat content in a one-stop way. However, Multi-Echo 3D GRE sequence is a work-in-progress project which cannot be used clinically and support only certain MR scanners. IDEAL-IQ sequence meets the similar limitation and can only be used on certain GE MR scanners, though IDEAL-IQ is a commercial sequence. Comparing with the above, commercial dual-echo imaging sequences are the most common method in routine clinical application and can be found in almost all MR scanners. Yet the separate impacts of T1 and T2* relaxation effect in pancreatic fat quantification using dual-echo imaging at 3.0-T MR have not been reported in detail.

Hence, this study aimed to analyze in detail the separate impact of T1 and T2* relaxation effects and show how the corrections work, and explore the feasibility of simplified correction process in dual-echo imaging of pancreatic fat quantification at 3.0-T MR for convenient clinical application.

Material and Methods

Participants

This study was approved by the local Ethics Committee. Written informed consent was obtained from all participants. The study included 21 adult participants diagnosed with NAFLD who were at high risk of pancreatic steatosis (19). Patients with moderate or heavy alcohol consumption; chronic liver disease other than NAFLD, such as viral liver disease; use of drugs associated with steatosis; exposure to other known hepatotoxins were excluded. Cases with hemosiderosis were excluded. There were 21 participants (16 men, 5 women; mean age = 42.5 ± 6.2 years; age range = 31–55 years).

¹H-MR spectroscopy and dual-echo imaging

MRI and ¹H-MRS were performed using a 3.0-T MR scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany). A breath-hold single-voxel point-resolved spectroscopy (PRESS) sequence was performed for ¹H-MRS. The scan parameters were: TR/TE = 3000/12, 24, 36, 48, 76 ms; averages = 1; acquisition time = 18 s in a single breath-hold. Standard shimming was performed during free breathing and water suppression was not applied. T1 effect was eliminated using TR 3000 ms, while five TEs were used to estimate the T2 values of water and fat for T2 effect correction. Routine clinical breath-hold turbo-spin-echo T2-weighted (T2W) sequence was used for locating 10 × 10 × 10 mm volume of interest (VOI). VOI was placed at the center of the pancreatic head by a radiologist with four years of experience in abdominal MRI diagnosis, avoiding interference by peripheral fat tissue and the duct as much as possible (Fig. 1).

Fig. 1.

MR and ¹H-MRS images from a 34-year-old man with pancreatic caput fat fraction of 9.8% by ¹H-MRS. (a, b) IP and OP images from a clinical routine spoiled gradient echo sequence. The ROI (white circle) was placed at the pancreatic caput. (c, d) Axial and coronal T2W images for the location of VOI in ¹H-MRS. The VOI (white square) was placed at the pancreatic caput. (e) Spectrum image of ¹H-MRS obtained at the first TE (12 ms) with water peak at 4.7 ppm and fat peak at 1.3 ppm. (f) Exponential fitting curves of water (above) and fat (below) for correcting T2 effect, generated from five TEs (12, 24, 36, 48, 76 ms).

A clinical routine spoiled gradient echo sequence was applied to obtain dual-echo images with the following parameters: TR = 80 ms; IP TE = 2.46 ms and OP TE = 1.23 ms; field of view (FOV) = 415 × 335 cm; matrix = 352 × 286; FA = 50°; slice thickness = 5 mm; scan time = 15 s in a single breath-hold. Automatic shimming was performed by the MR scanner.

A multi-echo spoiled gradient echo sequence was performed to estimated equal T2* of fat and water: TR = 9.15 ms; TEs = 1.05, 2.46, 3.69, 4.92, 6.15, and 7.38 ms; FOV = 420 × 315 cm; matrix = 160 × 95; FA = 4°; slice thickness = 3.5 mm; scan time = 13 s in a single breath-hold. Automatic shimming was performed by the MR scanner.

All the patients were trained before MR scan to make sure the maximum consistency of inspired air in each breath, and the scans were performed at the end of inspiration.

¹H-MR spectroscopy data analysis

Post-processing, performed on the Siemens Syngo-imaging workstation, generally involved noise filtering, apodization, baseline adjustment, phase correction, signal fitting of the peaks within the acquired spectra, and integration of area under water and fat peaks at 4.7 ppm and 1.3 ppm, respectively. Pancreatic fat fraction (FF) of ¹H-MRS was automatically calculated by extrapolating fat and water integrals for TE = 0 ms using an exponential fitting of points acquired with single voxel spectroscopy at five different TEs. Calculation method was defined by Eq. 1:

{FF}_{MRS} = \frac{A_{0_fat}}{A_{0_fat} + A_{0_water}}

(1)

A_{0_fat} and A_{0_water} were the T2 corrected integral areas under fat and water peaks at TE = 0 ms, respectively.

Dual-echo imaging data analysis

For dual-echo images, the region of interest (ROI) was placed at the slice according to and in the center of the VOI in ¹H-MRS sequence as much as possible (Fig. 1). ROI signal intensity was measured on the Siemens Syngo-imaging workstation by a radiologist with five years of experience in abdominal MRI diagnosis. Eq. 2 is the classical Dixon (dual-echo imaging) equation for fat quantification without correction. Eqs. 3–5 were derived for correcting T1, T2*, and T1 + T2* relaxation effects in dual-echo imaging according to Cassidy et al. (11):

{FF}_{no correction} = \frac{{SI}_{ip} - {SI}_{op}}{2 * {SI}_{ip}}

(2)

{FF}_{T 1 correction} = \frac{a * ({SI}_{ip} - {SI}_{op})}{a * ({SI}_{ip} - {SI}_{op}) + b * ({SI}_{ip} + {SI}_{op})}

(3)

{FF}_{T 2^{*} correction} = \frac{{SI}_{ip} * e^{y} - {SI}_{op} * e^{x}}{2 * {SI}_{ip} * e^{y}}

(4)

{FF}_{T 1 & T 2^{*} correction} = \frac{a * ({SI}_{ip} * e^{y} - {SI}_{op} * e^{x})}{{SI}_{ip} * e^{y} * (a + b) + {SI}_{op} * e^{x} * (b - a)}

(5)

SI_ip and SI_op were the signal intensities of ROI at IP and OP images. a, b, e^x, and e^y were defined as follows:

a, b = sin α * \frac{1 - e^{- TR / T 1_{w, f}}}{1 - cos α * e^{- TR / T 1_{w, f}}}

(6)

e^{x, y} = e^{- {TE}_{ip, op} / T 2^{*}}

(7)

where α indicated the FA, T1_{w, f} indicated T1 values of pancreatic water (725 ms) and pancreatic fat (382 ms), respectively, from published values (20). T2* was estimated for each participant using a non-linear fitting algorithm by the gradient multi-echo sequence mentioned above. Mean T2* value (T2*_(m)) was calculated from all participants to substitute for T2* in Eq. 7, then FF_{T2* correction(m)} and FF_{T1+T2* correction(m)} based on Eqs. 4 and 5 were calculated using T2*(m).

Statistical analysis

Statistical analyses were performed with the SPSS statistical software (version 18.0, SPSS Inc., Chicago, IL, USA). The measurement data were expressed as means ± standard deviation (SD) and were accurate to one decimal place. Pancreatic fat fraction quantified by ¹H-MRS (FF_MRS) was used as the reference standard. Pearson’s correlation analysis and linear regression were performed to estimate the relationship between fat fractions measured by dual-echo imaging (FF_{no correction}, FF_{T1 correction}, FF_{T2* correction}, FF_{T1+T2* correction,} FF_{T2* correction(m),} FF_{T1+T2* correction(m)}) and ¹H-MRS. The Bland–Altman plot was used to evaluate the agreement between fat fractions measured with dual-echo imaging and ¹H-MRS by calculating the mean difference, the 95% limits of agreement, and the largest absolute value of difference (LAVD) (the lower the absolute mean difference and LAVD, the higher the agreement). Box-whisker plot was used to reveal the distribution of T2*. P < 0.05 was considered to be statistically significant.

Results

All 21 individuals successfully completed the examinations. All the data are revealed in Table 1. Mean fat fractions measured by dual-echo imaging were 5.2 ± 2.0%, 4.7 ± 1.7%, 6.9 ± 1.9%, 6.2 ± 1.7%, 7.0 ± 2.1%, and 6.3 ± 1.8% for FF_{no correction}, FF_{T1 correction}, FF_{T2* correction}, FF_{T1+T2* correction}, FF_{T2* correction(m)}, and FF_{T1+T2* correction(m)}, respectively, and 5.9 ± 1.7% for ¹H-MRS. Significant positive correlations (all P < 0.01) were found between the fat fraction measured by each dual-echo method and ¹H-MRS, with the correlation coefficient (r) of 0.86, 0.91, 0.92, 0.95, 0.91, and 0.93, respectively. T2*_(m) was 32.6 ± 4.0 ms (range = 25.3–41.2 ms). Box-whisker plot showed a good consistency of T2* with absence of any outlier (Fig. 2). Linear regressions (trend lines) of fat fraction measured by each correction method of dual-echo imaging to that of ¹H-MRS are demonstrated in Fig. 3. As compared to ¹H-MRS, the Bland–Altman plot showed the different mean differences, 95% limits of agreement, and LAVD for dual-echo imaging with different corrections, in which FF_{T1+T2* correction} showed the best agreement, while FF_{T1 correction} showed the worst agreement. The performance for all the dual-echo methods except FF_{T1+T2* correction(m)} in 95% limits of agreement were the same that only one value exceeded limit line lightly and there was no exceeded value for FF_{T1+T2* correction(m)} (Fig. 4).

Table 1.

Mean FFs measured by dual-echo imaging and ¹H-MRS. Correlation coefficients (r), mean differences, 95% limits of agreements, and largest absolute value of difference (LAVD) between FFs by dual-echo imaging with different correction and ¹H-MRS.

n = 21	FF_{no correction}	FF_T1correction	FF_{T2*correction}	FF_{T1+T2*correction}	FF_{T2* correction(m)}	FF_{T1+T2*correction(m)}	FF_MRS
Mean FF	5.2 ± 2.0%	4.7 ± 1.7%	6.9 ± 1.9%	6.2 ± 1.7%	7.0 ± 2.1%	6.3 ± 1.8%	5.9 ± 1.7%
R	0.86	0.91	0.92	0.95	0.91	0.93	–
Mean difference	–0.70	–1.21	1.04	0.31	1.15	0.43	–
95% limits of agreement	1.15/–2.56	0.17/–2.60	2.58/–0.51	1.47/–0.85	2.87/–0.57	1.76/–0.89	–
LAVD	2.0	2.5	2.4	1.0	2.6	1.3	–

Fig. 2.

Box-whisker plot of T2* values from all 21 participants, with the absence of an outlier.

Fig. 3.

Scatter diagram and trend-lines from linear regression showing the distribution trend of pancreatic fat fractions measured by dual-echo imaging with different corrections. T1 + T2* correction line shows the best correlation with the identity line (¹H-MRS), while no correction shows the worst.

Fig. 4.

The Bland–Altman plots of FFs by dual-echo imaging with different corrections as compared to that of ¹H-MRS. T1 + T2* correction shows the best agreement with ¹H-MRS, while T1 correction shows the worst.

Discussion

In the clinical setting of dual-echo imaging, a large FA is always chosen for obtaining high-contrast images, which are helpful for diagnosis. However, a large FA causes relative signal amplification of fat as compared to water, thus enhancing T1 bias in fat quantification (10). Using a small FA can eliminate T1 bias to a large extent, which does not require T1 values for correction. However, such a method is sensitive to image noise (residual bias) (14). For example, using an FA of 5° will reduce the signal by 46% and 33% from the maximum amplitudes of fat and water, respectively, which may make it inaccurate for fat quantification at extremely low and high fat contents.

Given that T1 mapping is not included in routine clinical application (21), we used the estimated T1 values from published reports to correct T1 bias by a large FA (50°). After T1 correction, the correlation coefficient increased from 0.86 to 0.91, which demonstrated the benefit of T1 correction. Notably, the agreement became weak in Bland–Altman analysis (mean difference = 0.70–1.21 and LAVD = 2.0–2.5), which may result from using previously reported T1 values instead of separately measuring T1 values for water and fat, as the T1 values vary in different chemical environments (22). Also, the agreement reduction could be amplified at high fat content and in some disease states, particularly in the presence of iron or copper deposition (22,23). Leporq et al. found that in cases of very little fat content in the liver, the fat proton density might be very low as compared to water proton density and the T1 correction may become irrelevant (24). Similar result was observed in our study (Fig. 3), in which the lower regression lines of FF_{T1 correction} and FF_{no correction} were very close, indicating the irrelevant difference at very low fat content.

Fat quantification with dual-echo imaging relies on the signal difference between OP and IP images. This technique attributes all signal intensity loss between OP and IP images to phase interference effects between fat and water signals and ignores any contribution of T2* effect to signal loss, which may lead to errors of quantification (11). For example, if the OP echo is acquired before the IP echo, T2* decay and fat interference have opposing effects in the IP image resulting in underestimation of fat. However, if the OP is acquired after the IP echo, T2* decay leads to signal intensity loss on the OP image resulting in overestimation of fat. In our study, OP echo was acquired before IP echo and the underestimation of fat was observed consequently (Fig. 3), whereby the fat fraction increased markedly after T2* correction, especially at low fat content where the T2* decay effect may have completely offset the fat interference. We recommend the T2* correction in pancreatic fat quantification of dual-echo imaging, especially at low fat content.

In our study, as compared to the outcome of T1 correction, a similar decreasing trend of agreement was observed by Bland–Altman analysis between the fat fraction measured by ¹H-MRS and dual-echo imaging after T2*correction (mean difference increased from 0.70 to 1.04 and LAVD from 2.0 to 2.4). This could perhaps be from using a single T2*, which assumed the same T2* for both fat and water, although the T2* of water and fat are different. Chebrolu et al. demonstrated that errors in the liver fat quantification using a single T2* correction method can exceed 20% when T2* values of water and fat are very short (iron concentrations), particularly at higher fat fraction (16). Our study showed similar results wherein the regression line of FF_{T2* correction}, with increasing fat fraction, gradually moved away from the identity line.

By using the mean value of measured T2* (32.6 ms) for simplifying the correction process, an optimal performance was observed in which only a minor decrease occurred in the correlation coefficient between the fat fraction measured by ¹H-MRS and dual-echo imaging with T2* correction from 0.92 to 0.91 and T1 + T2* correction from 0.95 to 0.93, based on the consistency of measured T2* in Fig. 2 (no outlier) and minor decline of agreement from Bland–Altman analysis (Table 1). By using T2* value of 32.6 ms, instead of using additional multi-echo sequence to measure a single T2* or separate T2* of fat and water, we could simplify the process of T2* correction in pancreatic fat quantification of dual-echo imaging, which will facilitate convenient clinical application.

Our study showed that fat fraction measured by dual-echo imaging with T1 + T2* corrections had the best correlation and agreement with that of ¹H-MRS (r = 0.95, mean difference = 0.31), which met our expectation and was in accordance with previous studies (10,25 –27). However, fat fraction measured by dual-echo imaging without correction showed the worst correlation (r = 0.86) as compared to ¹H-MRS. Interestingly, the agreement between fat fraction measured by dual-echo imaging without correction and ¹H-MRS was even better than that of T2* correction and T1 correction (mean difference = –0.70 vs. 1.04 and –1.21, LAVD = 2.0 vs. 2.4 and 2.5). Perhaps the counterbalance between T1 and T2* relaxation effects, which conversely increased and decreased fat fraction, contributed to the improved agreement. Similarly, Joe et al. (28) estimated hepatic fat fraction of potential liver donors using dual-echo imaging without correction and showed that the correlation coefficient was 0.868 as compared to histopathology.

Our study had several limitations. First, the sample size was small and did not include any patient with obvious pancreatic iron deposition and large FF range, which makes our findings less applicable to diseases such as thalassemia and patients with large FF range. Due to the complex environment and small volume of the pancreas, we chose PRESS MRS sequence for higher signal-to-noise ratio rather than STEAM (stimulated echo acquisition mode), which could be affected by J-coupling effect (29). Moreover, the VOIs for MRS and ROIs for dual-echo imaging in our study were certainly different, which may lead to bias in the comparing of fat quantification. At last, dual-echo imaging can quantify the fat fraction only in the range of 0–50%. Although pancreatic fat fractions >50% are rare, they may occur.

In conclusion, although the fat fraction estimated by dual-echo imaging with corrections for both T1 and T2* represented the best correlation and agreement as compared to that of ¹H-MRS, the no correction dual-echo imaging also worked very well and may benefit from the offsetting between T1 and T2* relaxation effects, in which the prior OP echo played a key role. For simplifying the correction process of dual-echo imaging, using T1 values of 725 ms and 382 ms along with a single T2* value of 32.6 ms for pancreatic water and fat at 3.0-T MR could be recommended in pancreatic fat quantification for T1 and T2* correction in the absence of obvious iron overload in the pancreas.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Muoio

Newgard

. Obesity-related derangements in metabolic regulation. Biochemistry 2006; 75: 367–401.

Heilbronn

Smith

Ravussin

. Failure of fat cell proliferation, mitochondrial function and fat oxidation results in ectopic fat storage, insulin resistance and type II diabetes mellitus. Int J Obes Relat Metab Disord 2004; 28(Suppl. 4): S12–S21.

Kim

Cho

et al.

Quantitative assessment of pancreatic fat by using unenhanced CT: pathologic correlation and clinical implications. Radiology 2014; 271: 104–112.

Campbell-Tofte

Hansen

et al.

Increased lipids in non-lipogenic tissues are indicators of the severity of type 2 diabetes in mice. Prostaglandins Leukot Essent Fatty Acids 2007; 76: 9–18.

Tushuizen

Bunck

Pouwels

et al.

Pancreatic fat content and beta-cell function in men with and without type 2 diabetes. Diabetes Care 2007; 30: 2916–2921.

Thomas

Hamilton

Patel

et al.

Hepatic triglyceride content and its relation to body adiposity: a magnetic resonance imaging and proton magnetic resonance spectroscopy study. Gut 2005; 54: 122–127.

Holalkere

Kambadakone

et al.

Imaging-based quantification of hepatic fat: methods and clinical applications. Radiographics 2009; 29: 1253–1277.

Wieckowska

McCullough

Feldstein

. Noninvasive diagnosis and monitoring of nonalcoholic steatohepatitis: present and future. Hepatology 2007; 46: 582–589.

Liska

Dufour

Zern

et al.

Interethnic differences in muscle, liver and abdominal fat partitioning in obese adolescents. PLoS One 2007; 2: e569–e569.

10.

Kim

Nayak

et al.

Comparison of fat-water MRI and single-voxel MRS in the assessment of hepatic and pancreatic fat fractions in humans. Obesity (Silver Spring) 2010; 18: 841–847.

11.

Cassidy

Yokoo

Aganovic

et al.

Fatty liver disease: MR imaging techniques for the detection and quantification of liver steatosis. Radiographics 2009; 29: 231–260.

12.

Taouli

Ehman

Reeder

. Advanced MRI methods for assessment of chronic liver disease. Am J Roentgenol 2009; 193: 14–27.

13.

Guiu

Petit

Loffroy

et al.

Quantification of liver fat content: comparison of triple-echo chemical shift gradient-echo imaging and in vivo proton MR spectroscopy. Radiology 2009; 250: 95–102.

14.

Liu

McKenzie

et al.

Fat quantification with IDEAL gradient echo imaging: correction of bias from T(1) and noise. Magn Reson Med 2007; 58: 354–364.

15.

Bydder

Yokoo

Hamilton

et al.

Relaxation effects in the quantification of fat using gradient echo imaging. Magn Reson Imaging 2008; 26: 347–359.

16.

Chebrolu

Hines

et al.

Independent estimation of T*2 for water and fat for improved accuracy of fat quantification. Magn Reson Med 2010; 63: 849–857.

17.

O’Regan

Callaghan

Wylezinska-Arridge

et al.

Liver fat content and T2*: simultaneous measurement by using breath-hold multiecho MR imaging at 3.0 T–feasibility. Radiology 2008; 247: 550–557.

18.

McKenzie

Shimakawa

et al.

Multiecho reconstruction for simultaneous water-fat decomposition and T2* estimation. J Magn Reson Imaging 2007; 26: 1153–1161.

19.

Patel

Peterson

Lin

et al.

Insulin resistance increases MRI-estimated pancreatic fat in nonalcoholic fatty liver disease and normal controls. Gastroenterol Res Pract 2013; 2013: 498296–498296.

20.

De Bazelaire

Duhamel

Rofsky

et al.

MR imaging relaxation times of abdominal and pelvic tissues measured in vivo at 3.0 T: preliminary results. Radiology 2004; 230: 652–659.

21.

Deoni

Rutt

Peters

. Rapid combined T1 and T2 mapping using gradient recalled acquisition in the steady state. Magn Reson Med 2003; 49: 515–526.

22.

Das

Ray

. Wilson's disease: an update. Nat Clin Pract Neurol 2006; 2: 482–493.

23.

George

Goldwurm

MacDonald

et al.

Increased hepatic iron concentration in nonalcoholic steatohepatitis is associated with increased fibrosis. Gastroenterology 1998; 114: 311–318.

24.

Leporq

Ratiney

Pilleul

et al.

Liver fat volume fraction quantification with fat and water T1 and T 2* estimation and accounting for NMR multiple components in patients with chronic liver disease at 1.5 and 3.0 T. Eur Radiol 2013; 23: 2175–2186.

25.

Reeder

Cruite

Hamilton

et al.

Quantitative assessment of liver fat with magnetic resonance imaging and spectroscopy. J Magn Reson Imaging 2011; 34: spcone–spcone.

26.

Song

Yan

. Detection of hepatic and pancreatic fat infiltration in type II diabetes mellitus patients with IDEAL-Quant using 3.0T MR: comparison with single-voxel proton spectroscopy. Chin Med J (Engl) 2014; 127: 3548–3552.

27.

Sijens

Edens

Bakker

et al.

MRI-determined fat content of human liver, pancreas and kidney. World J Gastroenterol 2010; 16: 1993–1998.

28.

Joe

Lee

Kim

et al.

Quantification of hepatic macrosteatosis in living, related liver donors using T1-independent, T2*-corrected chemical shift MRI. J Magn Reson Imaging 2012; 36: 1124–1130.

29.

Hamilton

Middleton

Bydder

et al.

Effect of PRESS and STEAM sequences on magnetic resonance spectroscopic liver fat quantification. J Magn Reson Imaging 2009; 30: 145–152.

Quantification of pancreatic fat with dual-echo imaging at 3.0-T MR in clinical application: how do the corrections for T1 and T2* relaxation effect work and simplified correction strategy

Abstract

Background

Purpose

Material and Methods

Results

Conclusion

Keywords

Introduction

Material and Methods

Participants

1H-MR spectroscopy and dual-echo imaging

1H-MR spectroscopy data analysis

Dual-echo imaging data analysis

Statistical analysis

Results

Discussion

Footnotes

Declaration of Conflicting Interests

Funding

References

¹H-MR spectroscopy and dual-echo imaging

¹H-MR spectroscopy data analysis