Original Research n Gastrointestinal
Imaging
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Hepatic Steatosis in Living Liver Donor Candidates: Preoperative Assessment by Using Breath-hold Triple-Echo MR Imaging and 1H MR Spectroscopy1 Inpyeong Hwang, MD Jeong Min Lee, MD Kyoung Bun Lee, MD Jeong Hee Yoon, MD Berthold Kiefer, MD Joon Koo Han, MD Byung Ihn Choi, MD
1
From the Department of Radiology (I.H., J.M.L., J.H.Y., J.K.H., B.I.C.), Institute of Radiation Medicine (J.M.L., J.K.H., B.I.C.), and Department of Pathology (K.B.L.), Seoul National University Hospital, 28 Yeongon-dong, Jongnogu, Seoul 110-744, Korea; and Siemens Healthcare, Erlangen, Germany (B.K.). Received April 11, 2013; revision requested May 22; revision received October 19; accepted November 12; final version accepted December 9. Address correspondence to J.M.L. (e-mail: [email protected]).
Purpose:
To evaluate the diagnostic performance of both breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation magnetic resonance (MR) imaging (triple-echo imaging) and high-speed T2-corrected multiecho hydrogen 1 (1H) MR spectroscopy in the assessment of macrovesicular hepatic steatosis in living liver donor candidates by using histologic assessment as a reference standard.
Materials and Methods:
The institutional review board approved this retrospective study with waiver of the need to obtain informed consent. One hundred eighty-two liver donor candidates who had undergone preoperative triple-echo imaging and singlevoxel (3 3 3 3 3 cm) MR spectroscopy performed with a 3.0-T imaging unit and who had also undergone histologic evaluation of macrovesicular steatosis were included in this study. In part 1 of the study (n = 84), the Pearson correlation coefficient was calculated. Receiver operating characteristic (ROC) curve analysis was performed to detect substantial (10%) macrovesicular steatosis. In part 2 of the study, with a different patient group (n = 98), diagnostic performance was evaluated by using the diagnostic cutoff values determined in part 1 of the study.
Results:
The correlation coefficients of triple-echo MR imaging and MR spectroscopy with macrovesicular steatosis were 0.886 and 0.887, respectively. The areas under the ROC curve for detection of substantial macrovesicular steatosis were 0.959 and 0.988, with cutoff values of 4.93% and 5.79%, respectively, and without a significant difference (P = .328). In the part 2 study group, sensitivity and specificity were 90.9% (10 of 11) and 86.2% (75 of 87) for triple-echo MR imaging and 90.9% (10 of 11) and 86.2% (75 of 87) for MR spectroscopy, respectively.
Conclusion:
Either breath-hold triple-echo MR imaging or MR spectroscopy can be used to detect substantial macrovesicular steatosis in living liver donor candidates. In the future, this may allow selective biopsy in candidates who are expected to have substantial macrovesicular steatosis on the basis of MR-based hepatic fat fraction. RSNA, 2014
q
RSNA, 2014
q
730
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
H
epatic steatosis is an important risk factor in liver surgery, and its incidence is thought to be approximately 25% in living liver donors (1). In living donor liver transplantation, 30% or greater macrovesicular hepatic steatosis (characterized by hepatocytes that contain large lipid droplets and show nuclei that are displaced to the peripheries of the cells) is considered to be a major risk factor for poor graft function or primary graft nonfunction (1,2). In addition, it also affects the safety of the procedure for the donor, because even mild (,30%) macrovesicular steatosis determines the recovery of the remnant liver in the donor (3–5). Although there is controversy regarding what degree of hepatic steatosis contraindicates living donor liver transplantation, some medical centers use 10% macrovesicular steatosis as a safe margin for living donor liver transplantation to ensure donor safety (6–8). Therefore, accurate diagnosis and quantification of macrovesicular steatosis in potential living liver donor candidates is important. In many liver transplantation centers, liver biopsy is considered the reference
Advances in Knowledge nn Findings at breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging and hydrogen 1(1H) MR spectroscopy were well correlated with the histologic degree of macrovesicular steatosis, with correlation coefficients of 0.886 (95% confidence interval [CI]: 0.829, 0.925) and 0.887 (95% CI: 0.831, 0.926), respectively. nn Each MR method could depict macrovesicular hepatic steatosis of 10% or greater, which is clinically substantial in transplantation, with a sensitivity of 90.9% (10 of 11) and a specificity of 86.2% (75 of 87) for triple-echo MR imaging (cutoff value, 4.93%) and a sensitivity of 90.9% (10 of 11) and a specificity of 86.2% (75 of 87) for 1H MR spectroscopy (cutoff value, 5.79%).
standard (9). Although the major biopsy complication rate is relatively low (,1%) (9), this invasive procedure should be avoided if detection of hepatic steatosis is possible noninvasively. In addition, liver biopsy has disadvantages, such as small sample volumes and interobserver variability (10). Although many medical centers advocate routine liver biopsy as part of donor evaluation, recently, in some living liver transplantation centers, biopsy has been used selectively for the evaluation of steatosis when an imaging modality shows findings of hepatic steatosis, to avoid unnecessary biopsy (11,12). However, there is no consensus as to which imaging modality should be used or regarding the threshold values that should be used to identify hepatic steatosis and to suggest which patients require biopsy. One report (13) suggests that when an unenhanced computed tomographic (CT) study shows moderate (.30%) hepatic steatosis, biopsy should be avoided and the candidate should be excluded from living liver donor transplantation. However, using an alternative approach, liver biopsy could be indicated only if cross-sectional imaging studies show hepatic steatosis. There have been several attempts to evaluate hepatic steatosis by using various imaging tools, such as
Hwang et al
ultrasonography (US) (14), unenhanced CT (13,15), dual-echo gradient-echo magnetic resonance (MR) imaging (16), and hydrogen 1 (1H) MR spectroscopy (17,18). Among cross-sectional imaging modalities, MR techniques can decompose the liver signal into its fat and water signal components and can enable more direct assessment of the amount of fat in the liver than can either CT or US (19). Several previous studies (20,21) have demonstrated that MR techniques such as T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging (triple-echo imaging) and high-speed T2corrected multiecho MR spectroscopy can be used to evaluate the hepatic fat fraction (HFF) within one breath hold (22). However, to our knowledge, few studies have focused on clinical practice to advocate selective liver biopsy in living liver donors. Therefore, the goal of the present study was to evaluate the diagnostic performance of both breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging and high-speed T2-corrected multiecho 1H MR spectroscopy for the assessment of macrovesicular hepatic steatosis in living liver donor candidates, with histologic assessment as the reference standard.
Implications for Patient Care nn Preoperative assessment of substantial (10%) macrovesicular hepatic steatosis in living liver donors could be sufficiently achieved by the use of either breath-hold T2*-corrected tripleecho spoiled gradient-echo water-fat separation MR imaging or high-speed T2-corrected multiecho 1H MR spectroscopy. nn In the future it may be possible to perform preoperative selective liver biopsy only in those potential donors who are suspected of having substantial (10%) macrovesicular hepatic steatosis at triple-echo MR imaging and MR spectroscopy, rather than routine liver biopsy in all living liver donor candidates.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Published online before print 10.1148/radiol.14130863 Content codes: Radiology 2014; 271:730–738 Abbreviations: AUC = area under the receiver operating characteristic curve CI = confidence interval HFF = hepatic fat fraction ROI = region of interest Author contributions: Guarantor of integrity of entire study, J.M.L.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, I.H., K.B.L., J.H.Y., J.K.H., B.I.C.; clinical studies, I.H., J.M.L., K.B.L., J.H.Y., J.K.H., B.I.C.; statistical analysis, I.H., J.M.L., K.B.L.; and manuscript editing, I.H., J.M.L., B.K., B.I.C. Conflicts of interest are listed at the end of this article.
731
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Materials and Methods Study Population Our institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived. The sequences used for fat fraction estimation were provided by Siemens Healthcare (Erlangen, Germany). B.K. is an employee of Siemens Healthcare. Other authors who had no relationship with industry (I.H., J.M.L., and J.H.Y.) performed the study and had control of all of the data and information. Between June 2011 and November 2012, 215 consecutive living liver donor candidates who had no underlying liver disease underwent liver MR imaging for preoperative evaluation. These donor candidates were divided into two groups for two independent investigations: (a) Part 1 of this study (n = 107), which was performed during the first 8 months of the study period to determine the diagnostic criteria for the detection of macrovesicular steatosis, and (b) part 2 of this study (n = 108), which was performed during the latter 10 months of the study period to validate these criteria. In part 1 of the study, after the exclusion of candidates who did not have liver biopsy specimens available (n = 21) or did not undergo MR spectroscopy (n = 2), 84 donor candidates were ultimately included. The reasons 21 candidates did not have biopsy specimens available were as follows: A cadaveric donor became available (n = 7), another donor who was included in our study was ultimately used (n = 2), the recipient died before transplantation (n = 2), it was decided that liver transplantation was not indicated (eg, the recipient had multiple hepatocellular carcinoma metastases, portal vein tumor thrombus, or had recovered from acute hepatitis) (n = 6), the donor candidate was lost to follow-up (n = 1), the donor candidate withdrew although he or she did not have liver disease (n = 2), and the donor candidate was found to have dyslipidemia at biochemical assay (n = 1). In part 2 of the study, among 108 732
Hwang et al
Figure 1
Figure 1: Flowchart of the study population.
candidates, 10 who did not have liver specimens were excluded (no assessment of macrovesicular steatosis, n = 1; no surgery, n = 9). The reasons nine of these candidates did not undergo surgery were as follows: A cadaveric donor became available (n = 1); the recipient was found to have multiple hepatocellular carcinoma metastases (n = 1); the donor candidate withdrew although he or she did not have liver disease (n = 5); and increased HFF at MR imaging, without confirmation at liver biopsy (n = 2). Of the latter two candidates, one had HFFs of 21.9% at triple-echo MR imaging and 23.7% at MR spectroscopy, and the other had HFFs of 10.3% at triple-echo MR imaging and 12.2% at MR spectroscopy. Finally, 98 liver donor candidates were included in part 2 of the study. A total of 182 candidates were included in our study. The patient selection process is summarized in Figure 1, and demographic and clinical characteristics of the donor candidates are shown in Table 1.
MR Imaging All donor candidates underwent liver MR imaging performed with a 3.0-T MR imaging unit (Magnetom Verio; Siemens Healthcare) and by using a 32-channel phased-array coil. The routine liver MR pulse sequences for living liver donor work-up included breathhold triple-echo imaging and MR
spectroscopy. These fat-quantification sequences were performed before the intravenous administration of gadolinium-based contrast medium (gadoxetic acid, Primovist; Bayer Schering Pharma, Berlin, Germany). Axial T2*-corrected triple-echo spoiled gradient-echo water-fat separation imaging (triple-echo imaging) was performed by using a three-dimensional gradient-echo sequence. The triple-echo data consisted of in-phase (360°), opposed-phase (540°), and inphase (1080°) images acquired with the following parameters: repetition time msec/echo times msec, 9.89/2.45, 3.67, 7.35; flip angle, 11°; matrix, 256 3 167 3 36; 6/8 partial Fourier sampling along the phase and section directions; interpolated section thickness, 3.5 mm (56 sections); number of signals acquired, one; bandwidth, 400 Hz/pixel; a monopolar readout gradient; and field of view, 380 3 327 mm. Parallel imaging (generalized autocalibrating partially parallel acquisitions; Siemens Healthcare) with an effective acceleration factor of 1.5 was applied in an in-plane, phase-encoding direction. The acquisition time was 22 seconds, which was sufficient to allow imaging during one breath hold. Voxel-wise correction of T2* decay was applied by using T2* values calculated from the signal intensities and the echo times in two in-phase images by using a linear fit in log space. By using the magnitude
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Table 1 Demographic Characteristics, Clinical Characteristics, and Mean Macrovesicular Steatosis Values in Study Population Parameter M:F ratio Age (y) Overall Male donor candidates Female donor candidates Body mass index (kg/m2) Overall Male donor candidates Female donor candidates Time interval between imaging and biopsy (d) Percentage of candidates with (prevalence of) 10% macrovesicular steatosis* Mean percentage of macrovesicular steatosis In candidates with macrovesicular steatosis , 10% In candidates with macrovesicular steatosis 10% Overall
Part 1 Study Group (n = 84)
Part 2 Study Group (n = 98)
Overall (n = 182)
62:22
66:32
128:54
33 (17–61) 30 (17–58) 39 (18–61)
33 (16–61) 32 (16–54) 37 (16–61)
24.1 (17.1–31.5) 24.4 (17.1–31.5) 23.1 (19.9–26.9) 13 (0–55) 9.5 (8/84)
2.03 6 2.05 (0–7)
33 (16–61) 31 (16–58) 38 (16–61)
23.7 (16.0–34.5) 24.0 (17.4–32.3) 23.0 (16.0–34.5) 13 (0–83)
23.9 (15.9–34.5) 24.2 (17.1–32.3) 23.0 (16.0–34.5) 13 (0–83)
11.2 (11/98)
10.4 (19/182)
2.06 6 2.10 (0–7)
2.04 6 2.07 (0–7)
16.9 6 6.51 (10–30)
15.18 6 8.15 (10–30)
15.89 6 7.36 (10–30)
3.45 6 5.15 (0–30)
3.53 6 5.30 (0–30)
3.49 6 5.22 (0–30)
Note.—Unless otherwise specified, data are means 6 standard deviations, with ranges in parentheses. * Data in parentheses are raw data.
images of T2*-corrected in-phase and opposed-phase images, two-point water-fat separation was performed to calculate the HFF. To reduce the T1 weighting effect, a small flip angle (11°) was used. The water-fat ambiguities were resolved by using unwrapped phase data (not included in the sequence output). The final sequence outputs consisted of original triple-echo images, water-only and fat-only images, a T2* map, and an HFF map.
mm2 (range, 266–411 mm2) were drawn in the right half of the liver (Couinaud segments V, VIII, and VII) (Fig 2). Large intrahepatic vessels, bile ducts, and space-occupying lesions were avoided when drawing the ROIs, with reference to the original triple-echo images. The investigator was blinded to the exact location of the biopsy site, and exact spatial correlation was not possible. No images had substantial artifacts that disrupted the ROI measurements.
MR Image Analysis One radiologist (I.H., with 3 years of experience in reading liver MR images) performed region of interest (ROI) measurements in the right half of the liver according to the HFF percentage map. On each series, 15 circular ROIs (three ROIs per section times five consecutive sections) of approximately 300
Hydrogen 1 MR Spectroscopic Data Acquisition and Measurement High-speed T2-corrected multiecho single-voxel 1H MR spectroscopy was performed by using a modified stimulated-echo acquisition sequence at echo times of 12, 24, 36, 48, and 72 msec. The single voxel (3 3 3 3 3 cm) was manually placed in the right liver lobe
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
in the dome area of Couinaud segment VII or VIII while avoiding large blood vessels, large bile ducts, and liver edges. The repetition time was 3000 msec, and the mixing time between the second and third radiofrequency pulses was 10 msec so as to minimize T1 weighting. The flip angle was 90°. Each acquisition was performed during one breath hold (15 seconds). The first point of the free induction decay signal of every single channel was used to perform coil combination; the magnitude of the first point was used as an estimate of the coil channel sensitivity, and its phase was used to phase every signal for a constructive combination. Filtering was performed by using a simple low-pass sliding window filter. A simple linear baseline correction was performed by computing the average of 200 signal points upfield. Each of the T2 values of water and fat was calculated separately, and T2 correction was applied for both the water and fat peaks to obtain an accurate HFF (22). Fat quantification was performed by extrapolating water and fat integrals for an echo time of 0 msec by using an exponential fit of the points acquired at five different echoes by automatic fitting of water (4.7 ppm) and the major fat peak of methylene (1.3 ppm) (Fig 2). For fitting of water and fat, the peaks we searched for were the maxima of the magnitude spectra in specific ranges where water and fat are expected. Starting from the water (or fat) peak, the algorithm looked for the point in the spectrum where the signal reached 10% of the maximum in the direction of the fat (or water) peak. The middle of the two points obtained defined one of the limits of the interval used for integration of the water and lipid signal. Two other points were predefined to the right and to the left of the spectrum and were used as the other limits of the integration interval. The HFF was calculated automatically and was displayed as a percentage in Digital Imaging and Communications in Medicine format.
Histologic Analysis The degree of macrovesicular steatosis was assessed by means of visual interpretation of the liver biopsy specimen. 733
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Hwang et al
Figure 2
Figure 2: Representative examples of ROI measurement on (a) axial HFF map. Three approximately 300-mm2 ROIs per section were drawn in the right half of the liver. (b) Top: representative spectrum at MR spectroscopy for echo time (TE) of 12 msec (as in a). Red line = fat peak, blue line = water peak. Bottom: plot of echo time versus signal intensity for water and fat extrapolated from top graph to an echo time of 0 msec. Red line = water, blue line = fat. a.u. = Arbitrary units, Chem. = chemical, S.I. = signal intensity.
In our medical center, only selected donor candidates undergo percutaneous liver biopsy—those with abnormal biochemical liver function test results (eg, elevation of the serum aminotransferase level) and/or evidence of hepatic steatosis in imaging studies (US, CT, or MR imaging). If the candidate decided to undergo preoperative liver biopsy during the donor work-up, it was performed twice percutaneously from the right lobe of the liver (Couinaud segment VIII or VI) with an 18-gauge biopsy gun (Acecut 18G; TSK Laboratory, Tochigi, Japan) and US guidance by clinically experienced radiologists. An intraoperative liver biopsy (sample size, approximately 1 cm2) was performed in all liver donors. If the candidate had undergone percutaneous liver biopsy before surgery, the assessment of macrovesicular steatosis in the percutaneous liver biopsy specimen was used. In part 1 of our study, nine candidates underwent preoperative percutaneous liver biopsy, and the remaining 75 734
underwent intraoperative biopsy only. In part 2 of our study, 19 candidates underwent preoperative percutaneous liver biopsy, and the remaining 79 underwent intraoperative biopsy only. The degree of macrovesicular steatosis, which is the percentage of hepatocytes that contain intracellular macrovesicular fat droplets, was assessed quantitatively with hematoxylin-eosin staining by one of two liver pathologists at our institute (K.B.L., with 9 years of clinical experience in assessing liver disease, and another pathologist, with 31 years of experience). We considered macrovesicular steatosis of 10% or greater to be substantial macrovesicular steatosis in identifying those candidates needing further evaluation or dietary intervention to achieve weight reduction before donation.
Statistical Analysis All statistical analyses were performed by using software (MedCalc 12.2; MedCalc Software, Mariakerke, Belgium).
In part 1 of the study, we calculated the Pearson correlation coefficient between each HFF obtained by using the MR techniques and the histologic degree of macrovesicular steatosis; linear regression was also performed. Receiver operating characteristic curve analysis was performed to evaluate diagnostic performance and to determine optimal cutoff values at triple-echo MR imaging and MR spectroscopy for identifying substantial macrovesicular steatosis among the living liver donor candidates. Optimal cutoff values were determined with at least 90% of specificity. In part 2 of the study, the criteria from part of the 1 study were applied to another living liver donor candidate population (the part 2 study group), and in the entire study population, sensitivity, specificity, and accuracy were calculated. To assess the regional heterogeneity of HFF, the average standard deviation of HFF from 15 ROIs on triple-echo images was calculated in the entire study population. P , .05
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
was considered to indicate a significant difference.
Results Histologic Assessment of HFF In our study population, there were 19 (10.4%) living donor candidates with substantial (10%) macrovesicular steatosis. The prevalence of substantial macrovesicular steatosis and the mean macrovesicular steatosis in both study groups are summarized in Table 1. There was no significant difference in the prevalence of substantial macrovesicular steatosis between the subgroups (P = .810). The maximum degree of macrovesicular steatosis seen in our study population was 30%. Part 1 of the Study Comparison of results of each MR technique with histologic findings.—The correlation coefficient for triple-echo imaging and macrovesicular steatosis was r of 0.886 (95% confidence interval [CI]: 0.829, 0.925), and that for the comparison between MR spectroscopy and macrovesicular steatosis was r of 0.887 (95% CI: 0.831, 0.926) (P = .976). Figure 3 shows the scatterplots for each MR-based HFF and macrovesicular steatosis. Excellent correlation was observed between triple-echo imaging and MR spectroscopy (r = 0.939; 95% CI: 0.907, 0.960) (Fig 3). Determination of cutoff values of each technique for detecting substantial macrovesicular steatosis.—The areas under the receiver operating characteristic curve (AUCs) for detecting substantial macrovesicular steatosis at triple-echo imaging and MR spectroscopy were 0.959 (95% CI: 0.892, 0.990) and 0.988 (95% CI: 0.936, 1.000), respectively, with no significant difference (P = .328). The corresponding optimal cutoff values were 4.93% and 5.79%, respectively (Table 2). Part 2 of the Study Validation of the diagnostic criteria and comparison with clinical decision to perform selective biopsy.—Table 2 summarizes the diagnostic performance of
each criterion in terms of identifying substantial macrovesicular steatosis in the part 2 study group, as well as in the total study population. There was one false-negative result for the detection of substantial macrovesicular steatosis in the part 2 study group (ie, the HFF was 4.3% at triple-echo MR imaging and 5.12% at MR spectroscopy, both methods thus indicating , 10% macrovesicular steatosis). However, the patient’s serum aminotransferase level was elevated, and the percutaneous biopsy revealed 10% macrovesicular steatosis. This candidate underwent surgery after confirmation of 10% macrovesicular steatosis at intraoperative biopsy.
Regional Heterogeneity of HFF at MR Imaging The average standard deviation of HFF in 15 ROIs in our entire study population was 0.52% (range, 0%–1.87%) at triple-echo MR imaging. The maximum standard deviation was 1.87%, and this result suggests that the regional hepatic fat distributions in our study population were relatively homogeneous. Discussion Our study results demonstrate that breath-hold MR–based HFF has a strong correlation with macrovesicular steatosis, although they have different scales. Macrovesicular steatosis is defined as the percentage of cells affected by macrovesicular fatty change, while triple-echo imaging– or MR spectroscopy–measured HFF is calculated from the signal intensities of protons in water and fat. Our correlation coefficient results are in good agreement with those of previous studies that used similar triple-echo techniques (r = 0.848–0.90) (20,23,24) or MR spectroscopy (r = 0.85–0.95) (18,23) and used histologic hepatic steatosis as a reference standard. Breath-hold MR spectroscopy requires only 15 seconds, and automatic calculation of the HFF was possible at our MR workstation. Although in the past, MR spectroscopy has been considered the most precise, noninvasive
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
technique for evaluating the HFF, it has not been widely used in clinical practice primarily because it requires a long acquisition time and complex postprocessing analysis that requires an expert to interpret the spectral peaks. We therefore believe that the short acquisition time we used and the ability to automatically calculate the HFF by using MR spectroscopy may increase its clinical value. In our study, triple-echo MR imaging also showed roughly equivalent results to those of MR spectroscopy in correlation with macrovesicular steatosis. Although we had no study participants with marked heterogeneous hepatic steatosis, there were well-known varieties of hepatic steatosis, such as focal steatosis, hypersteatosis, and fatty sparing (25). A published study (26) demonstrated that there was less correlation between HFF and histologic findings of steatosis for heterogeneous hepatic steatosis. Considering that triple-echo imaging makes it possible to perform whole-liver fat quantification within one breath hold, which would not be possible at MR spectroscopy or liver biopsy, it might be more beneficial for evaluating patients with uneven fatty infiltration. Therefore, triple-echo MR imaging could be used as an initial imaging technique for assessing the HFF distribution in the whole liver. Hepatic iron content is a major confounding factor that causes acceleration of T2* decay, leading to underestimation of HFF (19). Although we had T2* maps produced by triple-echo imaging, histologic evaluation of iron accumulation, which would have required additional histologic staining, was beyond the scope of our study. However, several previous studies (20,27) reported that T2* corrections are effective for correcting the confounding effects of iron and allow more accurate calculation of HFF than dual-echo gradient-echo imaging (23). In addition, our study was performed with a 3.0-T unit, on which T2* decay is accelerated. Because similar studies performed with 3.0-T MR imaging units (18,28,29) showed consistent results regarding the measurement of HFF, the effect of accelerated 735
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Hwang et al
Figure 3
Figure 3: Scatterplots show HFFs plotted against macrovesicular steatosis for (a) triple-echo MR imaging and (b) MR spectroscopy. In addition, (c) a scatterplot of HFFs at MR spectroscopy against those at triple-echo MR imaging was drawn for comparison of the two MR methods. Solid line = regression line, dashed line = 95% predictive line, gray line = line of equity, horizontal dot-and-dash line = cutoff value.
T2* decay due to hepatic iron deposition is thought to be negligible with robust T2* correction. In our study, the diagnostic performance of both triple-echo MR imaging and MR spectroscopy in the detection of macrovesicular steatosis of 10% or greater was excellent. Therefore, we believe that the use of each of the two techniques or of both techniques could potentially replace liver biopsy as part of the preoperative assessment of living liver donor candidates. The cutoff values for detection of substantial macrovesicular steatosis were 4.93% at triple-echo MR imaging and 5.79% at MR spectroscopy, which are 736
comparable to the 5.56% cutoff value in a previous study (30). A recently published study (6) reported that macrovesicular steatosis of 10% or greater was the second most common reason for donor disqualification, while the most common cause was donor reluctance. In other words, the evaluation of macrovesicular steatosis is the most important reason for preoperative liver biopsy, and our results suggest that triple-echo MR imaging and MR spectroscopy can be used as an initial evaluation tool to determine the necessity of selective preoperative liver biopsy in living donor liver transplantation. Further prospective clinical studies should
be performed to assess whether MRbased HFF can eliminate the clinical demand for liver biopsy and how it correlates with clinical outcomes, such as the primary liver nonfunction rate and the patient survival rate, in living donor liver transplantation. There were a few limitations to our study. First, because it was conducted retrospectively, there may have been a potential selection bias. The prevalence of macrovesicular steatosis of 10% or greater is similar to that in a previous study of a living liver donor candidate population in another Asian country (11). However, because we did not include any candidates who had macrovesicular steatosis of more than 30%, further studies should be performed to determine whether estimation of the HFF with our MR method could also be correct in patients with marked hepatic steatosis. Second, our MR technique did not consider all known confounding factors in the measurement of HFF at MR. To reduce the T1 weighting effect, we used a small flip angle (11°), although it might still have been large for a 3.0-T imaging unit. Another consideration was that neither of our methods
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
88.5 (161/182) 89.0 (162/182) 88.3 (144/163) 88.3 (144/163) 86.7 (85/98) 86.7 (85/98) 86.2 (75/87) 86.2 (75/87)
* Data in parentheses are 95% CIs.
Note.—Unless otherwise specified, data in parentheses are raw data.
4.93 5.79 0.959 (0.892, 0.990) 0.988 (0.936, 1.000) Triple-echo MR imaging MR spectroscopy
87.5 (7/8) 100 (8/8)
90.8 (69/76) 90.8 (69/76)
90.9 (10/11) 90.9 (10/11)
89.5 (17/19) 94.7 (18/19)
Accuracy (%) Specificity (%)
Entire Study Population
Sensitivity (%) Accuracy (%) Part 2 Study
Specificity (%) Sensitivity (%) Specificity (%) Sensitivity (%)
Study Part 1
Cutoff Level (%) AUC* Imaging Modality
Study Part 1 Diagnostic Imaging Performance and Cutoff Values and Results of Validation of Diagnostic Performance for the Detection of Substantial Macrovesicular Steatosis in Study Part 2
Table 2
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
considered the multipeak fat spectrum. These confounding factors limit the generalized application of our study results. Recently, proton-density fat fraction imaging that corrects all of the known confounding factors has come to be regarded as accurate, precise, and a possible alternative biomarker of hepatic steatosis (29,31–33), and it would be helpful in the future to further assess this technique. Another limitation was that the spatial correlation between the biopsy site and the measurement site of HFF was not considered because exact information regarding the surgical biopsy site was not available in our study. Fourth, all of the MR examinations in our study were performed by using the same MR unit (Verio; Siemens Healthcare), and therefore, it might not be possible to extrapolate our study results to those of an actual clinical situation in which multiple MR examinations would be performed by using various MR units (35). Finally, there were some candidates who had a relatively long interval (.2 weeks) between their MR examination and surgical biopsy. Considering that short-term dietary intervention could change the degree of hepatic steatosis (34), the long time interval could have caused the reduced correlation in our study. In conclusion, both breath-hold T2*corrected triple-echo spoiled gradientecho water-fat separation MR imaging and high-speed T2-corrected multiecho MR spectroscopy have 90.9% sensitivity for detecting substantial (10%) macrovesicular steatosis in living liver donor candidates. We believe that performing preoperative selective liver biopsy only in candidates in whom we expect substantial macrovesicular steatosis on the basis of findings at tripleecho MR imaging and MR spectroscopy could thus be advocated. Disclosures of Conflicts of Interest: I.H. No relevant conflicts of interest to disclose. J.M.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: has received grants and personal fees from Bayer Healthcare, grants and nonfinancial support from GE Healthcare, grants from Guerbe, grants from Dong Seo Medical, grants from RF Medical,
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
grants from Starmed, grants from Toshiba Healthcare, and nonfinancial support from Philips. Other relationships: none to disclose. K.B.L. No relevant conflicts of interest to disclose. J.H.Y. No relevant conflicts of interest to disclose. B.K. Financial activities related to the present article: is an employee of Siemens Healthcare. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. J.K.H. No relevant conflicts of interest to disclose. B.I.C. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: has received grants from Samsung Electronic. Other relationships: none to disclose.
References 1. Selzner M, Clavien PA. Fatty liver in liver transplantation and surgery. Semin Liver Dis 2001;21(1):105–113. 2. Perez-Daga JA, Santoyo J, Suárez MA, et al. Influence of degree of hepatic steatosis on graft function and postoperative complications of liver transplantation. Transplant Proc 2006;38(8):2468–2470. 3. Nagai S, Fujimoto Y, Kamei H, Nakamura T, Kiuchi T. Mild hepatic macrovesicular steatosis may be a risk factor for hyperbilirubinaemia in living liver donors following right hepatectomy. Br J Surg 2009;96(4):437– 444. 4. Cho JY, Suh KS, Lee HW, et al. Hepatic steatosis is associated with intrahepatic cholestasis and transient hyperbilirubinemia during regeneration after living donor liver transplantation. Transpl Int 2006;19(10):807–813. 5. Shin YH, Ko JS, Kim GS, et al. Impact of hepatic macrovesicular and microvesicular steatosis on the postoperative liver functions after right hepatectomy in living donors. Transplant Proc 2012;44(2):512–515. 6. Sharma A, Ashworth A, Behnke M, Cotterell A, Posner M, Fisher RA. Donor selection for adult-to-adult living donor liver transplantation: well begun is half done. Transplantation 2013;95(3):501–506. 7. Trotter JF, Campsen J, Bak T, et al. Outcomes of donor evaluations for adult-toadult right hepatic lobe living donor liver transplantation. Am J Transplant 2006;6(8): 1882–1889. 8. Valentín-Gamazo C, Malagó M, Karliova M, et al. Experience after the evaluation of 700 potential donors for living donor liver transplantation in a single center. Liver Transpl 2004;10(9):1087–1096. 9. Brandhagen D, Fidler J, Rosen C. Evaluation of the donor liver for living donor liver transplantation. Liver Transpl 2003;9(10 Suppl 2):S16–S28.
737
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
10. El-Badry AM, Breitenstein S, Jochum W, et al. Assessment of hepatic steatosis by expert pathologists: the end of a gold standard. Ann Surg 2009;250(5):691–697. 11. Yamashiki N, Sugawara Y, Tamura S, et al. Noninvasive estimation of hepatic steatosis in living liver donors: usefulness of visceral fat area measurement. Transplantation 2009; 88(4):575–581. 12. Ahn JS, Sinn DH, Gwak GY, et al. Steatosis among living liver donors without evidence of fatty liver on ultrasonography: potential implications for preoperative liver biopsy. Transplantation 2013;95(11):1404–1409. 13. Park SH, Kim PN, Kim KW, et al. Mac rovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006;239(1): 105–112. 14. Kim SH, Lee JM, Kim JH, et al. Appropriateness of a donor liver with respect to macrosteatosis: application of artificial neural networks to US images—initial experience. Radiology 2005;234(3):793–803. 15. Lee SW, Park SH, Kim KW, et al. Unen hanced CT for assessment of macrovesicular hepatic steatosis in living liver donors: comparison of visual grading with liver attenuation index. Radiology 2007;244(2):479–485. 16. Kim SH, Lee JM, Han JK, et al. Hepatic macrosteatosis: predicting appropriateness of liver donation by using MR imaging—correlation with histopathologic findings. Radiology 2006;240(1):116–129. 17. Hájek M, Dezortová M, Wagnerová D, et al. MR spectroscopy as a tool for in vivo determination of steatosis in liver transplant recipients. MAGMA 2011;24(5):297–304. 18. Koelblinger C, Krššák M, Maresch J, et al. Hepatic steatosis assessment with 1H-spectroscopy and chemical shift imaging at 3.0 T before hepatic surgery: reliable enough for making clinical decisions? Eur J Radiol 2012;81(11):2990–2995.
738
19. Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative assessment of liver fat with magnetic resonance imaging and spectroscopy. J Magn Reson Imaging 2011;34(4): 729–749. 20. Kühn JP, Evert M, Friedrich N, et al. Noninvasive quantification of hepatic fat content using three-echo Dixon magnetic resonance imaging with correction for T2* relaxation effects. Invest Radiol 2011;46(12):783–789. 21. Guiu B, Petit JM, Loffroy R, 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(1):95–102. 22. Pineda N, Sharma P, Xu Q, Hu X, Vos M, Martin DR. Measurement of hepatic lipid: high-speed T2-corrected multiecho acquisition at 1H MR spectroscopy—a rapid and accurate technique. Radiology 2009;252(2): 568–576. 23. Kang BK, Yu ES, Lee SS, et al. Hepatic fat quantification: a prospective comparison of magnetic resonance spectroscopy and analysis methods for chemical-shift gradient echo magnetic resonance imaging with histologic assessment as the reference standard. Invest Radiol 2012;47(6):368–375. 24. Kühn JP, Hernando D, Muñoz del Rio A, et al. Effect of multipeak spectral modeling of fat for liver iron and fat quantification: correlation of biopsy with MR imaging results. Radiology 2012;265(1):133–142. 25. Karcaaltincaba M, Akhan O. Imaging of hepatic steatosis and fatty sparing. Eur J Radiol 2007;61(1):33–43. 26. Noworolski SM, Lam MM, Merriman RB, Ferrell L, Qayyum A. Liver steatosis: concordance of MR imaging and MR spectroscopic data with histologic grade. Radiology 2012;264(1):88–96. 27. Lee SS, Lee Y, Kim N, et al. Hepatic fat quantification using chemical shift MR imaging and MR spectroscopy in the presence
Hwang et al
of hepatic iron deposition: validation in phantoms and in patients with chronic liver disease. J Magn Reson Imaging 2011;33(6): 1390–1398. 28. Yokoo T, Shiehmorteza M, Hamilton G, et al. Estimation of hepatic proton-density fat fraction by using MR imaging at 3.0 T. Radiology 2011;258(3):749–759. 29. Tang A, Tan J, Sun M, et al. Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis. Radiology 2013;267(2):422–431. 30. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab 2005; 288(2):E462–E468. 31. Idilman IS, Aniktar H, Idilman R, et al. Hepatic steatosis: quantification by proton density fat fraction with MR imaging versus liver biopsy. Radiology 2013;267(3):767– 775. 32. Joe E, Lee JM, Kim KW, et al. Quantifi cation of hepatic macrosteatosis in living, related liver donors using T1-independent, T2*-corrected chemical shift MRI. J Magn Reson Imaging 2012;36(5):1124–1130. 33. Meisamy S, Hines CD, Hamilton G, et al. Quantification of hepatic steatosis with T1-independent, T2-corrected MR imaging with spectral modeling of fat: blinded comparison with MR spectroscopy. Radiology 2011;258(3):767–775. 34. Nakamuta M, Morizono S, Soejima Y, et al. Short-term intensive treatment for donors with hepatic steatosis in living-donor liver transplantation. Transplantation 2005;80(5): 608–612. 35. Yoon JH, Lee JM, Woo HS, et al. Staging of hepatic fibrosis: comparison of magnetic resonance elastography and shear wave elastography in the same individuals. Korean J Radiol 2013;14(2):202–212.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
Imaging
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Hepatic Steatosis in Living Liver Donor Candidates: Preoperative Assessment by Using Breath-hold Triple-Echo MR Imaging and 1H MR Spectroscopy1 Inpyeong Hwang, MD Jeong Min Lee, MD Kyoung Bun Lee, MD Jeong Hee Yoon, MD Berthold Kiefer, MD Joon Koo Han, MD Byung Ihn Choi, MD
1
From the Department of Radiology (I.H., J.M.L., J.H.Y., J.K.H., B.I.C.), Institute of Radiation Medicine (J.M.L., J.K.H., B.I.C.), and Department of Pathology (K.B.L.), Seoul National University Hospital, 28 Yeongon-dong, Jongnogu, Seoul 110-744, Korea; and Siemens Healthcare, Erlangen, Germany (B.K.). Received April 11, 2013; revision requested May 22; revision received October 19; accepted November 12; final version accepted December 9. Address correspondence to J.M.L. (e-mail: [email protected]).
Purpose:
To evaluate the diagnostic performance of both breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation magnetic resonance (MR) imaging (triple-echo imaging) and high-speed T2-corrected multiecho hydrogen 1 (1H) MR spectroscopy in the assessment of macrovesicular hepatic steatosis in living liver donor candidates by using histologic assessment as a reference standard.
Materials and Methods:
The institutional review board approved this retrospective study with waiver of the need to obtain informed consent. One hundred eighty-two liver donor candidates who had undergone preoperative triple-echo imaging and singlevoxel (3 3 3 3 3 cm) MR spectroscopy performed with a 3.0-T imaging unit and who had also undergone histologic evaluation of macrovesicular steatosis were included in this study. In part 1 of the study (n = 84), the Pearson correlation coefficient was calculated. Receiver operating characteristic (ROC) curve analysis was performed to detect substantial (10%) macrovesicular steatosis. In part 2 of the study, with a different patient group (n = 98), diagnostic performance was evaluated by using the diagnostic cutoff values determined in part 1 of the study.
Results:
The correlation coefficients of triple-echo MR imaging and MR spectroscopy with macrovesicular steatosis were 0.886 and 0.887, respectively. The areas under the ROC curve for detection of substantial macrovesicular steatosis were 0.959 and 0.988, with cutoff values of 4.93% and 5.79%, respectively, and without a significant difference (P = .328). In the part 2 study group, sensitivity and specificity were 90.9% (10 of 11) and 86.2% (75 of 87) for triple-echo MR imaging and 90.9% (10 of 11) and 86.2% (75 of 87) for MR spectroscopy, respectively.
Conclusion:
Either breath-hold triple-echo MR imaging or MR spectroscopy can be used to detect substantial macrovesicular steatosis in living liver donor candidates. In the future, this may allow selective biopsy in candidates who are expected to have substantial macrovesicular steatosis on the basis of MR-based hepatic fat fraction. RSNA, 2014
q
RSNA, 2014
q
730
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
H
epatic steatosis is an important risk factor in liver surgery, and its incidence is thought to be approximately 25% in living liver donors (1). In living donor liver transplantation, 30% or greater macrovesicular hepatic steatosis (characterized by hepatocytes that contain large lipid droplets and show nuclei that are displaced to the peripheries of the cells) is considered to be a major risk factor for poor graft function or primary graft nonfunction (1,2). In addition, it also affects the safety of the procedure for the donor, because even mild (,30%) macrovesicular steatosis determines the recovery of the remnant liver in the donor (3–5). Although there is controversy regarding what degree of hepatic steatosis contraindicates living donor liver transplantation, some medical centers use 10% macrovesicular steatosis as a safe margin for living donor liver transplantation to ensure donor safety (6–8). Therefore, accurate diagnosis and quantification of macrovesicular steatosis in potential living liver donor candidates is important. In many liver transplantation centers, liver biopsy is considered the reference
Advances in Knowledge nn Findings at breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging and hydrogen 1(1H) MR spectroscopy were well correlated with the histologic degree of macrovesicular steatosis, with correlation coefficients of 0.886 (95% confidence interval [CI]: 0.829, 0.925) and 0.887 (95% CI: 0.831, 0.926), respectively. nn Each MR method could depict macrovesicular hepatic steatosis of 10% or greater, which is clinically substantial in transplantation, with a sensitivity of 90.9% (10 of 11) and a specificity of 86.2% (75 of 87) for triple-echo MR imaging (cutoff value, 4.93%) and a sensitivity of 90.9% (10 of 11) and a specificity of 86.2% (75 of 87) for 1H MR spectroscopy (cutoff value, 5.79%).
standard (9). Although the major biopsy complication rate is relatively low (,1%) (9), this invasive procedure should be avoided if detection of hepatic steatosis is possible noninvasively. In addition, liver biopsy has disadvantages, such as small sample volumes and interobserver variability (10). Although many medical centers advocate routine liver biopsy as part of donor evaluation, recently, in some living liver transplantation centers, biopsy has been used selectively for the evaluation of steatosis when an imaging modality shows findings of hepatic steatosis, to avoid unnecessary biopsy (11,12). However, there is no consensus as to which imaging modality should be used or regarding the threshold values that should be used to identify hepatic steatosis and to suggest which patients require biopsy. One report (13) suggests that when an unenhanced computed tomographic (CT) study shows moderate (.30%) hepatic steatosis, biopsy should be avoided and the candidate should be excluded from living liver donor transplantation. However, using an alternative approach, liver biopsy could be indicated only if cross-sectional imaging studies show hepatic steatosis. There have been several attempts to evaluate hepatic steatosis by using various imaging tools, such as
Hwang et al
ultrasonography (US) (14), unenhanced CT (13,15), dual-echo gradient-echo magnetic resonance (MR) imaging (16), and hydrogen 1 (1H) MR spectroscopy (17,18). Among cross-sectional imaging modalities, MR techniques can decompose the liver signal into its fat and water signal components and can enable more direct assessment of the amount of fat in the liver than can either CT or US (19). Several previous studies (20,21) have demonstrated that MR techniques such as T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging (triple-echo imaging) and high-speed T2corrected multiecho MR spectroscopy can be used to evaluate the hepatic fat fraction (HFF) within one breath hold (22). However, to our knowledge, few studies have focused on clinical practice to advocate selective liver biopsy in living liver donors. Therefore, the goal of the present study was to evaluate the diagnostic performance of both breath-hold T2*-corrected triple-echo spoiled gradient-echo water-fat separation MR imaging and high-speed T2-corrected multiecho 1H MR spectroscopy for the assessment of macrovesicular hepatic steatosis in living liver donor candidates, with histologic assessment as the reference standard.
Implications for Patient Care nn Preoperative assessment of substantial (10%) macrovesicular hepatic steatosis in living liver donors could be sufficiently achieved by the use of either breath-hold T2*-corrected tripleecho spoiled gradient-echo water-fat separation MR imaging or high-speed T2-corrected multiecho 1H MR spectroscopy. nn In the future it may be possible to perform preoperative selective liver biopsy only in those potential donors who are suspected of having substantial (10%) macrovesicular hepatic steatosis at triple-echo MR imaging and MR spectroscopy, rather than routine liver biopsy in all living liver donor candidates.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Published online before print 10.1148/radiol.14130863 Content codes: Radiology 2014; 271:730–738 Abbreviations: AUC = area under the receiver operating characteristic curve CI = confidence interval HFF = hepatic fat fraction ROI = region of interest Author contributions: Guarantor of integrity of entire study, J.M.L.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, I.H., K.B.L., J.H.Y., J.K.H., B.I.C.; clinical studies, I.H., J.M.L., K.B.L., J.H.Y., J.K.H., B.I.C.; statistical analysis, I.H., J.M.L., K.B.L.; and manuscript editing, I.H., J.M.L., B.K., B.I.C. Conflicts of interest are listed at the end of this article.
731
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Materials and Methods Study Population Our institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived. The sequences used for fat fraction estimation were provided by Siemens Healthcare (Erlangen, Germany). B.K. is an employee of Siemens Healthcare. Other authors who had no relationship with industry (I.H., J.M.L., and J.H.Y.) performed the study and had control of all of the data and information. Between June 2011 and November 2012, 215 consecutive living liver donor candidates who had no underlying liver disease underwent liver MR imaging for preoperative evaluation. These donor candidates were divided into two groups for two independent investigations: (a) Part 1 of this study (n = 107), which was performed during the first 8 months of the study period to determine the diagnostic criteria for the detection of macrovesicular steatosis, and (b) part 2 of this study (n = 108), which was performed during the latter 10 months of the study period to validate these criteria. In part 1 of the study, after the exclusion of candidates who did not have liver biopsy specimens available (n = 21) or did not undergo MR spectroscopy (n = 2), 84 donor candidates were ultimately included. The reasons 21 candidates did not have biopsy specimens available were as follows: A cadaveric donor became available (n = 7), another donor who was included in our study was ultimately used (n = 2), the recipient died before transplantation (n = 2), it was decided that liver transplantation was not indicated (eg, the recipient had multiple hepatocellular carcinoma metastases, portal vein tumor thrombus, or had recovered from acute hepatitis) (n = 6), the donor candidate was lost to follow-up (n = 1), the donor candidate withdrew although he or she did not have liver disease (n = 2), and the donor candidate was found to have dyslipidemia at biochemical assay (n = 1). In part 2 of the study, among 108 732
Hwang et al
Figure 1
Figure 1: Flowchart of the study population.
candidates, 10 who did not have liver specimens were excluded (no assessment of macrovesicular steatosis, n = 1; no surgery, n = 9). The reasons nine of these candidates did not undergo surgery were as follows: A cadaveric donor became available (n = 1); the recipient was found to have multiple hepatocellular carcinoma metastases (n = 1); the donor candidate withdrew although he or she did not have liver disease (n = 5); and increased HFF at MR imaging, without confirmation at liver biopsy (n = 2). Of the latter two candidates, one had HFFs of 21.9% at triple-echo MR imaging and 23.7% at MR spectroscopy, and the other had HFFs of 10.3% at triple-echo MR imaging and 12.2% at MR spectroscopy. Finally, 98 liver donor candidates were included in part 2 of the study. A total of 182 candidates were included in our study. The patient selection process is summarized in Figure 1, and demographic and clinical characteristics of the donor candidates are shown in Table 1.
MR Imaging All donor candidates underwent liver MR imaging performed with a 3.0-T MR imaging unit (Magnetom Verio; Siemens Healthcare) and by using a 32-channel phased-array coil. The routine liver MR pulse sequences for living liver donor work-up included breathhold triple-echo imaging and MR
spectroscopy. These fat-quantification sequences were performed before the intravenous administration of gadolinium-based contrast medium (gadoxetic acid, Primovist; Bayer Schering Pharma, Berlin, Germany). Axial T2*-corrected triple-echo spoiled gradient-echo water-fat separation imaging (triple-echo imaging) was performed by using a three-dimensional gradient-echo sequence. The triple-echo data consisted of in-phase (360°), opposed-phase (540°), and inphase (1080°) images acquired with the following parameters: repetition time msec/echo times msec, 9.89/2.45, 3.67, 7.35; flip angle, 11°; matrix, 256 3 167 3 36; 6/8 partial Fourier sampling along the phase and section directions; interpolated section thickness, 3.5 mm (56 sections); number of signals acquired, one; bandwidth, 400 Hz/pixel; a monopolar readout gradient; and field of view, 380 3 327 mm. Parallel imaging (generalized autocalibrating partially parallel acquisitions; Siemens Healthcare) with an effective acceleration factor of 1.5 was applied in an in-plane, phase-encoding direction. The acquisition time was 22 seconds, which was sufficient to allow imaging during one breath hold. Voxel-wise correction of T2* decay was applied by using T2* values calculated from the signal intensities and the echo times in two in-phase images by using a linear fit in log space. By using the magnitude
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Table 1 Demographic Characteristics, Clinical Characteristics, and Mean Macrovesicular Steatosis Values in Study Population Parameter M:F ratio Age (y) Overall Male donor candidates Female donor candidates Body mass index (kg/m2) Overall Male donor candidates Female donor candidates Time interval between imaging and biopsy (d) Percentage of candidates with (prevalence of) 10% macrovesicular steatosis* Mean percentage of macrovesicular steatosis In candidates with macrovesicular steatosis , 10% In candidates with macrovesicular steatosis 10% Overall
Part 1 Study Group (n = 84)
Part 2 Study Group (n = 98)
Overall (n = 182)
62:22
66:32
128:54
33 (17–61) 30 (17–58) 39 (18–61)
33 (16–61) 32 (16–54) 37 (16–61)
24.1 (17.1–31.5) 24.4 (17.1–31.5) 23.1 (19.9–26.9) 13 (0–55) 9.5 (8/84)
2.03 6 2.05 (0–7)
33 (16–61) 31 (16–58) 38 (16–61)
23.7 (16.0–34.5) 24.0 (17.4–32.3) 23.0 (16.0–34.5) 13 (0–83)
23.9 (15.9–34.5) 24.2 (17.1–32.3) 23.0 (16.0–34.5) 13 (0–83)
11.2 (11/98)
10.4 (19/182)
2.06 6 2.10 (0–7)
2.04 6 2.07 (0–7)
16.9 6 6.51 (10–30)
15.18 6 8.15 (10–30)
15.89 6 7.36 (10–30)
3.45 6 5.15 (0–30)
3.53 6 5.30 (0–30)
3.49 6 5.22 (0–30)
Note.—Unless otherwise specified, data are means 6 standard deviations, with ranges in parentheses. * Data in parentheses are raw data.
images of T2*-corrected in-phase and opposed-phase images, two-point water-fat separation was performed to calculate the HFF. To reduce the T1 weighting effect, a small flip angle (11°) was used. The water-fat ambiguities were resolved by using unwrapped phase data (not included in the sequence output). The final sequence outputs consisted of original triple-echo images, water-only and fat-only images, a T2* map, and an HFF map.
mm2 (range, 266–411 mm2) were drawn in the right half of the liver (Couinaud segments V, VIII, and VII) (Fig 2). Large intrahepatic vessels, bile ducts, and space-occupying lesions were avoided when drawing the ROIs, with reference to the original triple-echo images. The investigator was blinded to the exact location of the biopsy site, and exact spatial correlation was not possible. No images had substantial artifacts that disrupted the ROI measurements.
MR Image Analysis One radiologist (I.H., with 3 years of experience in reading liver MR images) performed region of interest (ROI) measurements in the right half of the liver according to the HFF percentage map. On each series, 15 circular ROIs (three ROIs per section times five consecutive sections) of approximately 300
Hydrogen 1 MR Spectroscopic Data Acquisition and Measurement High-speed T2-corrected multiecho single-voxel 1H MR spectroscopy was performed by using a modified stimulated-echo acquisition sequence at echo times of 12, 24, 36, 48, and 72 msec. The single voxel (3 3 3 3 3 cm) was manually placed in the right liver lobe
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
in the dome area of Couinaud segment VII or VIII while avoiding large blood vessels, large bile ducts, and liver edges. The repetition time was 3000 msec, and the mixing time between the second and third radiofrequency pulses was 10 msec so as to minimize T1 weighting. The flip angle was 90°. Each acquisition was performed during one breath hold (15 seconds). The first point of the free induction decay signal of every single channel was used to perform coil combination; the magnitude of the first point was used as an estimate of the coil channel sensitivity, and its phase was used to phase every signal for a constructive combination. Filtering was performed by using a simple low-pass sliding window filter. A simple linear baseline correction was performed by computing the average of 200 signal points upfield. Each of the T2 values of water and fat was calculated separately, and T2 correction was applied for both the water and fat peaks to obtain an accurate HFF (22). Fat quantification was performed by extrapolating water and fat integrals for an echo time of 0 msec by using an exponential fit of the points acquired at five different echoes by automatic fitting of water (4.7 ppm) and the major fat peak of methylene (1.3 ppm) (Fig 2). For fitting of water and fat, the peaks we searched for were the maxima of the magnitude spectra in specific ranges where water and fat are expected. Starting from the water (or fat) peak, the algorithm looked for the point in the spectrum where the signal reached 10% of the maximum in the direction of the fat (or water) peak. The middle of the two points obtained defined one of the limits of the interval used for integration of the water and lipid signal. Two other points were predefined to the right and to the left of the spectrum and were used as the other limits of the integration interval. The HFF was calculated automatically and was displayed as a percentage in Digital Imaging and Communications in Medicine format.
Histologic Analysis The degree of macrovesicular steatosis was assessed by means of visual interpretation of the liver biopsy specimen. 733
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Hwang et al
Figure 2
Figure 2: Representative examples of ROI measurement on (a) axial HFF map. Three approximately 300-mm2 ROIs per section were drawn in the right half of the liver. (b) Top: representative spectrum at MR spectroscopy for echo time (TE) of 12 msec (as in a). Red line = fat peak, blue line = water peak. Bottom: plot of echo time versus signal intensity for water and fat extrapolated from top graph to an echo time of 0 msec. Red line = water, blue line = fat. a.u. = Arbitrary units, Chem. = chemical, S.I. = signal intensity.
In our medical center, only selected donor candidates undergo percutaneous liver biopsy—those with abnormal biochemical liver function test results (eg, elevation of the serum aminotransferase level) and/or evidence of hepatic steatosis in imaging studies (US, CT, or MR imaging). If the candidate decided to undergo preoperative liver biopsy during the donor work-up, it was performed twice percutaneously from the right lobe of the liver (Couinaud segment VIII or VI) with an 18-gauge biopsy gun (Acecut 18G; TSK Laboratory, Tochigi, Japan) and US guidance by clinically experienced radiologists. An intraoperative liver biopsy (sample size, approximately 1 cm2) was performed in all liver donors. If the candidate had undergone percutaneous liver biopsy before surgery, the assessment of macrovesicular steatosis in the percutaneous liver biopsy specimen was used. In part 1 of our study, nine candidates underwent preoperative percutaneous liver biopsy, and the remaining 75 734
underwent intraoperative biopsy only. In part 2 of our study, 19 candidates underwent preoperative percutaneous liver biopsy, and the remaining 79 underwent intraoperative biopsy only. The degree of macrovesicular steatosis, which is the percentage of hepatocytes that contain intracellular macrovesicular fat droplets, was assessed quantitatively with hematoxylin-eosin staining by one of two liver pathologists at our institute (K.B.L., with 9 years of clinical experience in assessing liver disease, and another pathologist, with 31 years of experience). We considered macrovesicular steatosis of 10% or greater to be substantial macrovesicular steatosis in identifying those candidates needing further evaluation or dietary intervention to achieve weight reduction before donation.
Statistical Analysis All statistical analyses were performed by using software (MedCalc 12.2; MedCalc Software, Mariakerke, Belgium).
In part 1 of the study, we calculated the Pearson correlation coefficient between each HFF obtained by using the MR techniques and the histologic degree of macrovesicular steatosis; linear regression was also performed. Receiver operating characteristic curve analysis was performed to evaluate diagnostic performance and to determine optimal cutoff values at triple-echo MR imaging and MR spectroscopy for identifying substantial macrovesicular steatosis among the living liver donor candidates. Optimal cutoff values were determined with at least 90% of specificity. In part 2 of the study, the criteria from part of the 1 study were applied to another living liver donor candidate population (the part 2 study group), and in the entire study population, sensitivity, specificity, and accuracy were calculated. To assess the regional heterogeneity of HFF, the average standard deviation of HFF from 15 ROIs on triple-echo images was calculated in the entire study population. P , .05
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
was considered to indicate a significant difference.
Results Histologic Assessment of HFF In our study population, there were 19 (10.4%) living donor candidates with substantial (10%) macrovesicular steatosis. The prevalence of substantial macrovesicular steatosis and the mean macrovesicular steatosis in both study groups are summarized in Table 1. There was no significant difference in the prevalence of substantial macrovesicular steatosis between the subgroups (P = .810). The maximum degree of macrovesicular steatosis seen in our study population was 30%. Part 1 of the Study Comparison of results of each MR technique with histologic findings.—The correlation coefficient for triple-echo imaging and macrovesicular steatosis was r of 0.886 (95% confidence interval [CI]: 0.829, 0.925), and that for the comparison between MR spectroscopy and macrovesicular steatosis was r of 0.887 (95% CI: 0.831, 0.926) (P = .976). Figure 3 shows the scatterplots for each MR-based HFF and macrovesicular steatosis. Excellent correlation was observed between triple-echo imaging and MR spectroscopy (r = 0.939; 95% CI: 0.907, 0.960) (Fig 3). Determination of cutoff values of each technique for detecting substantial macrovesicular steatosis.—The areas under the receiver operating characteristic curve (AUCs) for detecting substantial macrovesicular steatosis at triple-echo imaging and MR spectroscopy were 0.959 (95% CI: 0.892, 0.990) and 0.988 (95% CI: 0.936, 1.000), respectively, with no significant difference (P = .328). The corresponding optimal cutoff values were 4.93% and 5.79%, respectively (Table 2). Part 2 of the Study Validation of the diagnostic criteria and comparison with clinical decision to perform selective biopsy.—Table 2 summarizes the diagnostic performance of
each criterion in terms of identifying substantial macrovesicular steatosis in the part 2 study group, as well as in the total study population. There was one false-negative result for the detection of substantial macrovesicular steatosis in the part 2 study group (ie, the HFF was 4.3% at triple-echo MR imaging and 5.12% at MR spectroscopy, both methods thus indicating , 10% macrovesicular steatosis). However, the patient’s serum aminotransferase level was elevated, and the percutaneous biopsy revealed 10% macrovesicular steatosis. This candidate underwent surgery after confirmation of 10% macrovesicular steatosis at intraoperative biopsy.
Regional Heterogeneity of HFF at MR Imaging The average standard deviation of HFF in 15 ROIs in our entire study population was 0.52% (range, 0%–1.87%) at triple-echo MR imaging. The maximum standard deviation was 1.87%, and this result suggests that the regional hepatic fat distributions in our study population were relatively homogeneous. Discussion Our study results demonstrate that breath-hold MR–based HFF has a strong correlation with macrovesicular steatosis, although they have different scales. Macrovesicular steatosis is defined as the percentage of cells affected by macrovesicular fatty change, while triple-echo imaging– or MR spectroscopy–measured HFF is calculated from the signal intensities of protons in water and fat. Our correlation coefficient results are in good agreement with those of previous studies that used similar triple-echo techniques (r = 0.848–0.90) (20,23,24) or MR spectroscopy (r = 0.85–0.95) (18,23) and used histologic hepatic steatosis as a reference standard. Breath-hold MR spectroscopy requires only 15 seconds, and automatic calculation of the HFF was possible at our MR workstation. Although in the past, MR spectroscopy has been considered the most precise, noninvasive
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
technique for evaluating the HFF, it has not been widely used in clinical practice primarily because it requires a long acquisition time and complex postprocessing analysis that requires an expert to interpret the spectral peaks. We therefore believe that the short acquisition time we used and the ability to automatically calculate the HFF by using MR spectroscopy may increase its clinical value. In our study, triple-echo MR imaging also showed roughly equivalent results to those of MR spectroscopy in correlation with macrovesicular steatosis. Although we had no study participants with marked heterogeneous hepatic steatosis, there were well-known varieties of hepatic steatosis, such as focal steatosis, hypersteatosis, and fatty sparing (25). A published study (26) demonstrated that there was less correlation between HFF and histologic findings of steatosis for heterogeneous hepatic steatosis. Considering that triple-echo imaging makes it possible to perform whole-liver fat quantification within one breath hold, which would not be possible at MR spectroscopy or liver biopsy, it might be more beneficial for evaluating patients with uneven fatty infiltration. Therefore, triple-echo MR imaging could be used as an initial imaging technique for assessing the HFF distribution in the whole liver. Hepatic iron content is a major confounding factor that causes acceleration of T2* decay, leading to underestimation of HFF (19). Although we had T2* maps produced by triple-echo imaging, histologic evaluation of iron accumulation, which would have required additional histologic staining, was beyond the scope of our study. However, several previous studies (20,27) reported that T2* corrections are effective for correcting the confounding effects of iron and allow more accurate calculation of HFF than dual-echo gradient-echo imaging (23). In addition, our study was performed with a 3.0-T unit, on which T2* decay is accelerated. Because similar studies performed with 3.0-T MR imaging units (18,28,29) showed consistent results regarding the measurement of HFF, the effect of accelerated 735
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
Hwang et al
Figure 3
Figure 3: Scatterplots show HFFs plotted against macrovesicular steatosis for (a) triple-echo MR imaging and (b) MR spectroscopy. In addition, (c) a scatterplot of HFFs at MR spectroscopy against those at triple-echo MR imaging was drawn for comparison of the two MR methods. Solid line = regression line, dashed line = 95% predictive line, gray line = line of equity, horizontal dot-and-dash line = cutoff value.
T2* decay due to hepatic iron deposition is thought to be negligible with robust T2* correction. In our study, the diagnostic performance of both triple-echo MR imaging and MR spectroscopy in the detection of macrovesicular steatosis of 10% or greater was excellent. Therefore, we believe that the use of each of the two techniques or of both techniques could potentially replace liver biopsy as part of the preoperative assessment of living liver donor candidates. The cutoff values for detection of substantial macrovesicular steatosis were 4.93% at triple-echo MR imaging and 5.79% at MR spectroscopy, which are 736
comparable to the 5.56% cutoff value in a previous study (30). A recently published study (6) reported that macrovesicular steatosis of 10% or greater was the second most common reason for donor disqualification, while the most common cause was donor reluctance. In other words, the evaluation of macrovesicular steatosis is the most important reason for preoperative liver biopsy, and our results suggest that triple-echo MR imaging and MR spectroscopy can be used as an initial evaluation tool to determine the necessity of selective preoperative liver biopsy in living donor liver transplantation. Further prospective clinical studies should
be performed to assess whether MRbased HFF can eliminate the clinical demand for liver biopsy and how it correlates with clinical outcomes, such as the primary liver nonfunction rate and the patient survival rate, in living donor liver transplantation. There were a few limitations to our study. First, because it was conducted retrospectively, there may have been a potential selection bias. The prevalence of macrovesicular steatosis of 10% or greater is similar to that in a previous study of a living liver donor candidate population in another Asian country (11). However, because we did not include any candidates who had macrovesicular steatosis of more than 30%, further studies should be performed to determine whether estimation of the HFF with our MR method could also be correct in patients with marked hepatic steatosis. Second, our MR technique did not consider all known confounding factors in the measurement of HFF at MR. To reduce the T1 weighting effect, we used a small flip angle (11°), although it might still have been large for a 3.0-T imaging unit. Another consideration was that neither of our methods
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
88.5 (161/182) 89.0 (162/182) 88.3 (144/163) 88.3 (144/163) 86.7 (85/98) 86.7 (85/98) 86.2 (75/87) 86.2 (75/87)
* Data in parentheses are 95% CIs.
Note.—Unless otherwise specified, data in parentheses are raw data.
4.93 5.79 0.959 (0.892, 0.990) 0.988 (0.936, 1.000) Triple-echo MR imaging MR spectroscopy
87.5 (7/8) 100 (8/8)
90.8 (69/76) 90.8 (69/76)
90.9 (10/11) 90.9 (10/11)
89.5 (17/19) 94.7 (18/19)
Accuracy (%) Specificity (%)
Entire Study Population
Sensitivity (%) Accuracy (%) Part 2 Study
Specificity (%) Sensitivity (%) Specificity (%) Sensitivity (%)
Study Part 1
Cutoff Level (%) AUC* Imaging Modality
Study Part 1 Diagnostic Imaging Performance and Cutoff Values and Results of Validation of Diagnostic Performance for the Detection of Substantial Macrovesicular Steatosis in Study Part 2
Table 2
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
considered the multipeak fat spectrum. These confounding factors limit the generalized application of our study results. Recently, proton-density fat fraction imaging that corrects all of the known confounding factors has come to be regarded as accurate, precise, and a possible alternative biomarker of hepatic steatosis (29,31–33), and it would be helpful in the future to further assess this technique. Another limitation was that the spatial correlation between the biopsy site and the measurement site of HFF was not considered because exact information regarding the surgical biopsy site was not available in our study. Fourth, all of the MR examinations in our study were performed by using the same MR unit (Verio; Siemens Healthcare), and therefore, it might not be possible to extrapolate our study results to those of an actual clinical situation in which multiple MR examinations would be performed by using various MR units (35). Finally, there were some candidates who had a relatively long interval (.2 weeks) between their MR examination and surgical biopsy. Considering that short-term dietary intervention could change the degree of hepatic steatosis (34), the long time interval could have caused the reduced correlation in our study. In conclusion, both breath-hold T2*corrected triple-echo spoiled gradientecho water-fat separation MR imaging and high-speed T2-corrected multiecho MR spectroscopy have 90.9% sensitivity for detecting substantial (10%) macrovesicular steatosis in living liver donor candidates. We believe that performing preoperative selective liver biopsy only in candidates in whom we expect substantial macrovesicular steatosis on the basis of findings at tripleecho MR imaging and MR spectroscopy could thus be advocated. Disclosures of Conflicts of Interest: I.H. No relevant conflicts of interest to disclose. J.M.L. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: has received grants and personal fees from Bayer Healthcare, grants and nonfinancial support from GE Healthcare, grants from Guerbe, grants from Dong Seo Medical, grants from RF Medical,
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Hwang et al
grants from Starmed, grants from Toshiba Healthcare, and nonfinancial support from Philips. Other relationships: none to disclose. K.B.L. No relevant conflicts of interest to disclose. J.H.Y. No relevant conflicts of interest to disclose. B.K. Financial activities related to the present article: is an employee of Siemens Healthcare. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. J.K.H. No relevant conflicts of interest to disclose. B.I.C. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: has received grants from Samsung Electronic. Other relationships: none to disclose.
References 1. Selzner M, Clavien PA. Fatty liver in liver transplantation and surgery. Semin Liver Dis 2001;21(1):105–113. 2. Perez-Daga JA, Santoyo J, Suárez MA, et al. Influence of degree of hepatic steatosis on graft function and postoperative complications of liver transplantation. Transplant Proc 2006;38(8):2468–2470. 3. Nagai S, Fujimoto Y, Kamei H, Nakamura T, Kiuchi T. Mild hepatic macrovesicular steatosis may be a risk factor for hyperbilirubinaemia in living liver donors following right hepatectomy. Br J Surg 2009;96(4):437– 444. 4. Cho JY, Suh KS, Lee HW, et al. Hepatic steatosis is associated with intrahepatic cholestasis and transient hyperbilirubinemia during regeneration after living donor liver transplantation. Transpl Int 2006;19(10):807–813. 5. Shin YH, Ko JS, Kim GS, et al. Impact of hepatic macrovesicular and microvesicular steatosis on the postoperative liver functions after right hepatectomy in living donors. Transplant Proc 2012;44(2):512–515. 6. Sharma A, Ashworth A, Behnke M, Cotterell A, Posner M, Fisher RA. Donor selection for adult-to-adult living donor liver transplantation: well begun is half done. Transplantation 2013;95(3):501–506. 7. Trotter JF, Campsen J, Bak T, et al. Outcomes of donor evaluations for adult-toadult right hepatic lobe living donor liver transplantation. Am J Transplant 2006;6(8): 1882–1889. 8. Valentín-Gamazo C, Malagó M, Karliova M, et al. Experience after the evaluation of 700 potential donors for living donor liver transplantation in a single center. Liver Transpl 2004;10(9):1087–1096. 9. Brandhagen D, Fidler J, Rosen C. Evaluation of the donor liver for living donor liver transplantation. Liver Transpl 2003;9(10 Suppl 2):S16–S28.
737
GASTROINTESTINAL IMAGING: Assessment of Living Liver Donor Candidates with MR Imaging and MR Spectroscopy
10. El-Badry AM, Breitenstein S, Jochum W, et al. Assessment of hepatic steatosis by expert pathologists: the end of a gold standard. Ann Surg 2009;250(5):691–697. 11. Yamashiki N, Sugawara Y, Tamura S, et al. Noninvasive estimation of hepatic steatosis in living liver donors: usefulness of visceral fat area measurement. Transplantation 2009; 88(4):575–581. 12. Ahn JS, Sinn DH, Gwak GY, et al. Steatosis among living liver donors without evidence of fatty liver on ultrasonography: potential implications for preoperative liver biopsy. Transplantation 2013;95(11):1404–1409. 13. Park SH, Kim PN, Kim KW, et al. Mac rovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006;239(1): 105–112. 14. Kim SH, Lee JM, Kim JH, et al. Appropriateness of a donor liver with respect to macrosteatosis: application of artificial neural networks to US images—initial experience. Radiology 2005;234(3):793–803. 15. Lee SW, Park SH, Kim KW, et al. Unen hanced CT for assessment of macrovesicular hepatic steatosis in living liver donors: comparison of visual grading with liver attenuation index. Radiology 2007;244(2):479–485. 16. Kim SH, Lee JM, Han JK, et al. Hepatic macrosteatosis: predicting appropriateness of liver donation by using MR imaging—correlation with histopathologic findings. Radiology 2006;240(1):116–129. 17. Hájek M, Dezortová M, Wagnerová D, et al. MR spectroscopy as a tool for in vivo determination of steatosis in liver transplant recipients. MAGMA 2011;24(5):297–304. 18. Koelblinger C, Krššák M, Maresch J, et al. Hepatic steatosis assessment with 1H-spectroscopy and chemical shift imaging at 3.0 T before hepatic surgery: reliable enough for making clinical decisions? Eur J Radiol 2012;81(11):2990–2995.
738
19. Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative assessment of liver fat with magnetic resonance imaging and spectroscopy. J Magn Reson Imaging 2011;34(4): 729–749. 20. Kühn JP, Evert M, Friedrich N, et al. Noninvasive quantification of hepatic fat content using three-echo Dixon magnetic resonance imaging with correction for T2* relaxation effects. Invest Radiol 2011;46(12):783–789. 21. Guiu B, Petit JM, Loffroy R, 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(1):95–102. 22. Pineda N, Sharma P, Xu Q, Hu X, Vos M, Martin DR. Measurement of hepatic lipid: high-speed T2-corrected multiecho acquisition at 1H MR spectroscopy—a rapid and accurate technique. Radiology 2009;252(2): 568–576. 23. Kang BK, Yu ES, Lee SS, et al. Hepatic fat quantification: a prospective comparison of magnetic resonance spectroscopy and analysis methods for chemical-shift gradient echo magnetic resonance imaging with histologic assessment as the reference standard. Invest Radiol 2012;47(6):368–375. 24. Kühn JP, Hernando D, Muñoz del Rio A, et al. Effect of multipeak spectral modeling of fat for liver iron and fat quantification: correlation of biopsy with MR imaging results. Radiology 2012;265(1):133–142. 25. Karcaaltincaba M, Akhan O. Imaging of hepatic steatosis and fatty sparing. Eur J Radiol 2007;61(1):33–43. 26. Noworolski SM, Lam MM, Merriman RB, Ferrell L, Qayyum A. Liver steatosis: concordance of MR imaging and MR spectroscopic data with histologic grade. Radiology 2012;264(1):88–96. 27. Lee SS, Lee Y, Kim N, et al. Hepatic fat quantification using chemical shift MR imaging and MR spectroscopy in the presence
Hwang et al
of hepatic iron deposition: validation in phantoms and in patients with chronic liver disease. J Magn Reson Imaging 2011;33(6): 1390–1398. 28. Yokoo T, Shiehmorteza M, Hamilton G, et al. Estimation of hepatic proton-density fat fraction by using MR imaging at 3.0 T. Radiology 2011;258(3):749–759. 29. Tang A, Tan J, Sun M, et al. Nonalcoholic fatty liver disease: MR imaging of liver proton density fat fraction to assess hepatic steatosis. Radiology 2013;267(2):422–431. 30. Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab 2005; 288(2):E462–E468. 31. Idilman IS, Aniktar H, Idilman R, et al. Hepatic steatosis: quantification by proton density fat fraction with MR imaging versus liver biopsy. Radiology 2013;267(3):767– 775. 32. Joe E, Lee JM, Kim KW, et al. Quantifi cation of hepatic macrosteatosis in living, related liver donors using T1-independent, T2*-corrected chemical shift MRI. J Magn Reson Imaging 2012;36(5):1124–1130. 33. Meisamy S, Hines CD, Hamilton G, et al. Quantification of hepatic steatosis with T1-independent, T2-corrected MR imaging with spectral modeling of fat: blinded comparison with MR spectroscopy. Radiology 2011;258(3):767–775. 34. Nakamuta M, Morizono S, Soejima Y, et al. Short-term intensive treatment for donors with hepatic steatosis in living-donor liver transplantation. Transplantation 2005;80(5): 608–612. 35. Yoon JH, Lee JM, Woo HS, et al. Staging of hepatic fibrosis: comparison of magnetic resonance elastography and shear wave elastography in the same individuals. Korean J Radiol 2013;14(2):202–212.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
