- Email: [email protected]
Magnetic Resonance Spectroscopy in Living-Donor Liver Transplantation
Magnetic Resonance Spectroscopy in Living-Donor Liver Transplantation H.-J. Chianga, W.-P. Changa, H.-W. Chianga, M.Z. Lazoa, T.-Y. Chena, H.-Y. Oua, L.L.-C. Tsanga, T.-L. Huanga, C.-L. Chenb, and Y.-F. Chenga,* a
Department of Diagnostic Radiology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Niao-Sung, Kaosiung, Taiwan; and the bDepartment of Surgery, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Niao-Sung, Kaosiung, Taiwan
ABSTRACT Objective. The purpose of this study is to evaluate the accuracy rate of the one breathhold single voxel hydrogen-1 magnetic resonance spectroscopy (MRS) in comparison with intraoperative biopsy for liver fat quantification in living-donor liver transplantation. Materials and Methods. A total of 80 living liver donors participated in this study. Each patient underwent both MRS and intraoperative biopsy for evaluation of liver fatty content. MRS was performed using 1.5-T magnetic resonance imaging and placed in segments 2e4, 5e8, and left lateral segment for each donor. Accuracy was assessed through receiver operating characteristic curve analysis. Sensitivity and specificity of MRS fat fractions were also calculated. Results. Eighty living-donor liver transplantation donors were enrolled in this study. There was no fatty liver in 59 subjects (73.8%), 5% to 10% fatty liver in 17 subjects, 11% to 15% fatty liver in 3 subjects, and >16% fatty liver in 1 subject. MRS fat fraction showed excellent parameters to predict between normal liver and fatty liver groups (1.85% 0.98, 8.13% 3.52, respectively; P < .0001). Linear regression between MRS fat fraction and pathology grading showed high correlation (R2 ¼ 0.7092). Pearson correlation revealed high correlation between MRS and pathology results (r ¼ 0.936), poor correlation between body mass index and pathology results (r ¼ 0.390). The sensitivity and specificity for detection of liver steatosis in MRS fat fraction were 95.2% and 98.3%, respectively. Conclusion. 1H MRS fat fraction is a highly precise and accurate method in quantification of hepatic steatosis for the living donor and can be finished in a single breath-hold.
L
IVER transplantation is the best treatment modality for end-stage liver disease [1]. Because of the shortage of deceased donors, living-donor liver transplantation (LDLT) has become a primary treatment method. The most important ethical concern about LDLT is donor safety, as it has both surgical and health risks. Hepatic steatosis quantification is critical for donor selection in LDLT because graft steatosis is associated with an increased risk of complications after liver transplantation for both donor and recipient [2]. Hepatic steatosis, which is a common finding in living liver donors, not only influences the outcome of liver transplantation for the recipient but it also affects the recovery of the living donor after partial hepatectomy [3]. Therefore, many noninvasive imaging analysis methods are
used to quantify hepatic steatosis for preoperative livingdonor evaluation [4].
This work was supported by a grant from the Ministry of Health and Welfare, Taiwan (MOHW104-TDU-B-212-124-004 to Chen CL), and health and welfare surcharge of tobacco products. and grant CMRPG8C0011 from the Chang Gung Memorial Hospital research grant, Taiwan, and a Chang Gung Medical Foundation Institutional Review Board. Taiwan approval has been obtained (101-3673B). *Address correspondence to Yu-Fan Cheng, MD, Department of Diagnostic Radiology, Kaohsiung Chang Gung Memorial Hospital, 123 Ta-Pei Road, Niao-Sung, Kaosiung 833, Taiwan. E-mail: [email protected]
ª 2016 Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710
0041-1345/16 http://dx.doi.org/10.1016/j.transproceed.2015.10.068
Transplantation Proceedings, 48, 1003e1006 (2016)
1003
1004
Medical imaging has progressed rapidly in recent years. Several noninvasive methods have been developed for quantification of fat content, one of those is acoustic structure quantification of ultrasonography, which allows detection of hepatic steatosis that is 10% or greater in living liver donors [5]. This new technique has increased sensitivity in quantification of hepatic steatosis [6]; but there is still need for more precise quantification for living donor selection. Dual energy computed tomography (CT) fat analysis provides a noninvasive technique for identifying hepatic steatosis and has strongly correlated with histopathologic results [7]. However, radiation exposure has become a major problem in CT. Another limitation in CT imaging is that occasionally, pathologic abnormalities of the hepatic parenchyma are encountered such as iron overload that can influence the hepatic attenuation [8], and lead to masking or underestimation of hepatic steatosis in CT images. Magnetic resonance imaging is a highly sensitive tool for detection and characterization of fatty infiltration of liver. Magnetic resonance spectroscopy (MRS) has been used for detection and quantification of fatty infiltration in liver in recent years. MRS is one of the most accurate noninvasive techniques in assessment of hepatic steatosis. In this study, the investigators describe the 1-breath-hold stimulated echo acquisition mode (STEAM) hydrogen-1 MRS for quantification of hepatic steatosis; results are compared with liver biopsy results in the living liver donor. MATERIALS AND METHODS This prospective study was institutional review board approved by the human research committee of our institution.
Patients From March 2013 to May 2014, a total of 80 living donors (35 males and 45 females) underwent both pretransplantation MRS and intra-operative liver biopsy for liver fat quantification. Demographic characteristics and body mass index (BMI) were recorded. The mean age for all patients was 30.24 7.7 years (range 18 years to 47 years). The BMI ranged from 16.4 kg/m2 to 33.3 kg/m2 with a mean value of 23.4 kg/m2 and standard deviation of 4.1 kg/m2.
Image Acquisition Technique All the 1H MRSs were performed on a 1.5-T MR scanner (Discovery 450; GE Healthcare, Milwaukee, Wisconsin, United States). A body coil was used for signal excitation and an 8-channel body-phased array coil was used for signal reception. Breath-hold single-voxel MRS was acquired using the STEAM sequence. Voxel size was 2.0 cm3 2.0 cm3 2.0 cm3. The breath hold T1-weighted in/out phase images were used to locate the voxels placed in segments 2e4, 5e8, and left lateral segment for each donor; visible blood vessels or bile duct structures were avoided. Acquisition parameters for MRS are with a repetition time of 2,500 ms, an echo time of 12 ms, 2 next, 4 number of scan in each acquisition. In all cases, the quality of the shimming obtained in the voxel was controlled by the spectral line width (full width of half
CHIANG, CHANG, CHIANG ET AL maximum in Hz) of the unsuppressed water, obtained by the automated optimization sequence before scanning. No water suppression was applied to calculate the fat fraction of the liver. Total scan time was 21 seconds with a single breath-hold. 1
H MRS Post-processing
To estimate the liver fat fraction, all magnetic resonance spectra were analyzed with the spectral analysis program (SAGE 7.0; GE Healthcare) and performed with the same operator who has received spectrum analysis training. Post-processing steps include 8-channel signal combination, apodization, zero filling, Fourier transform, automated phase correction, and Marquardt curve fitting. The lipid signal peak was defined at 1.3 ppm, and the water was at 4.7 ppm. The signal fat fraction can then be given as the fat integral signal divided by the integral of the water and fat peaks areas.
Donor Biopsy Zero-hour biopsy specimens were obtained by wedge resection during surgery; sampling location was different for each subject, depending on the graft donated. Two independent pathologists performed histological grading of macrovesicular steatosis. For severity of fatty change and the presence of lobular inflammation, results were reported as a quantitative evaluation of the percentage of hepatocytes.
Statistical Analysis To determine the accuracy of the 1H MRS fat fraction, pathology grading was used as the gold standard. Hepatic steatosis from pathology reports in this study were divided into four groups. Group 1 included specimens normal to <5% fatty liver. Group 2 included specimens with 5% to 10% fatty liver. Group 3 included specimens with 11% to 15% fatty liver. Group 4 included specimens with >15% fatty liver. For statistical analysis, the pathology data were divided into two groups: normal (0 to <5% fatty change) and fatty liver (5% fatty change). The statistical analysis was based on independent paired Student t test. MRS fat fraction results were used to divide the subjects into two groups. Linear regression was used to detect the correlation between MRS fat fraction and pathology data. Receiver operating characteristic curve analysis was performed to analyze sensitivity and specificity for detection of steatosis in MRS fat fraction. Pearson correlation was used for the correlation among MRS fat fraction, BMI, and pathology data. All statistical analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, Illinois, United States). A P < .05 was considered significant.
RESULTS
Eighty LDLT donors were evaluated for fatty liver change by use of both MRS and intraoperative biopsy. Pathology results showed no fatty content in 73.8% (n ¼ 59), and fatty liver content in 26.2% (n ¼ 21). Of those with fatty liver, there were 21.2% with 5% to 10% fatty liver (n ¼ 17), 3.8% with 11% to 15% fatty liver (n ¼ 3), and 1.2% with >16% fatty liver (n ¼ 1). Independent paired Student t test showed that 1H MRS fat fraction was an excellent parameter to predict between normal and fatty liver groups (1.85% 0.98%, 8.13%
MAGNETIC RESONANCE SPECTROSCOPY
3.52%, P < .0001) in living donors, but no significant change in BMI (22.31 kg/m2 3.75, 26.45 kg/m2 3.65, P ¼ .863). Good linear regression was noted between 1H MRS and histopathologic results (R2 ¼ 0.7092). Pearson correlation showed a high correlation between 1H MRS and histopathologic results (r ¼ 0.936, P < .0001); however, a poor correlation is noted between BMI and histopathologic results (r ¼ 0.390, P < .0001). The sensitivity and specificity for detection of liver steatosis in 1H MRS were 95.2% and 98.3%, respectively (cutoff 4.965, area 0.999, P < .0001, 95% confidence interval 0.000e1.000) in receiver operating characteristic curve analysis. DISCUSSION
Because of the shortage of deceased donors for liver transplant, LDLT has become a widely accepted modality for treatment of end-stage liver disease in Taiwan, the top priority being donor safety [1]. Hepatic steatosis not only increases the risk of nonfunctioning or poorly functioning grafts, but it also increases the donor’s risk for postoperative complications. Regeneration of the donor’s remnant liver after major resection is impeded by hepatic steatosis. BMI has a positive correlation with increasing steatosis (r ¼ 0.390, P < .0001). Peng et al suggested that individuals with a BMI > 27.5 kg/m2 were most likely to show moderate steatosis and liver biopsy should be performed; on the other hand, donors with BMI < 23 kg/m2 were likely to show no or mild steatosis [9]. This study showed that among 17 donors with BMI > 27.5, 7 subjects (41.2%) exhibited no fatty change by biopsy. Among 43 donors with BMI < 23, 5 subjects (11.6%) demonstrated fatty change by biopsy. One donor with BMI ¼ 21.2 revealed 15% fatty change. In this regard, BMI value has no significant difference between the normal and fatty liver groups (22.31% 3.75, 26.45% 3.65, P ¼ .863). In our daily practice, the BMI was used as a reference in surveys for suitable donors in LDLT; in addition, BMI > 28 is also an indicator for liver biopsy. Unenhanced CT is being used for preoperative evaluation of fat content of donor liver. It has high diagnostic accuracy and high specificity for detecting more than 30% steatosis [10]. There is a paucity of literature to validate the utility of CT for quantitative assessment of liver steatosis and may not be as accurate for detection of mild steatosis [11,12]. Furthermore, radiation exposure makes unenhanced CT unsuitable for frequent monitoring of hepatic steatosis in potential living liver donors. In this study, MRS is used for quantification of liver fat fraction. The results showed high sensitivity and specificity in detection of steatosis, the cutoff value for detection of hepatic steatosis being 4.965% (close to 5%). MRS proves to be highly sensitive for detecting fatty liver change even in small amounts and this may be more suitable for preoperative evaluation. Biopsy is the gold standard for steatosis quantification, but it also shows sampling errors due to the heterogeneity of steatosis in the liver [13]. For this reason, three different
1005
voxels were placed to measure different segments of fat fraction to prevent sampling error and for greater precision of fat content quantification. The linear regression between 1 H MRS and histopathologic results in fatty group was 0.7092, taking into consideration the biopsy versus the MRS voxel locations which may not be the same; this variation may need more validation. The sensitivity and specificity achieved were 95.2% and 98.3%, respectively, for detection of liver steatosis in living donors. MRS allows the demonstration of relative tissue metabolite concentrations. Different metabolites have their own frequency, and the area under the signal peak in spectrum is equal to the compound concentration in the voxel [14]. Various magnetic resonance spectrum scan protocols can be used to detect fat content in the liver parenchyma. Spectroscopic imaging and chemical shift imaging use breathgating technique, which can do multi-voxel liver screen but may be time consuming (21 minutes) [15]. STEAM sequence uses a single-voxel technique to detect liver fat content. This sequence has less chance for overestimation of fat fraction due to J coupling effect in liver fat quantification and greater than point resolved spectroscopy sequence in liver fat quantification [16]. A single acquisition STEAM sequence was used in this study with a short scan time of 21 seconds and a single breath-holding technique to avoid respiratory movement problems and sampling bias during gating technique. Chemical shift imaging and iterative decomposition of water and fat with echo asymmetry and least squares estimation fat fraction imaging can obtain liver fat fraction for quantification of hepatic steatosis in the liver [17], but this is only the ratio not the fat concentration. MRS is the only noninvasive modality that can be used to determine the absolute liver fat concentration in liver. The concentration of fat in the voxel is equal to the area under the fat peak in spectrum; but the absolute quantitative value of liver fat concentration is more complex, which is not within the scope of this report because this study only discusses the fraction between the fat and water signal in spectroscopy in LDLT. 1 H MRS in vivo has limitations caused by complex spectrum analysis and technical issues. However, in this study, analysis of lipid peak (1.3 ppm) and water peak (4.7 ppm) to quantify fat fraction in the liver was performed with the use of commercialized post-processing software that can provide a semiautomatic spectrum analysis procedure. It can automatically calculate the area under the peak selected in the spectrum and instantly verify and modify the spectrum result. Therefore, the spectrum analysis has become easier to use and is faster for liver fat quantification. In conclusion, there are numerous benefits of the 1H MRS in quantification of fat infiltration of the liver for LDLT evaluation, including less systematic error, no radiation exposure (thus making it more reproducible for frequent monitoring when necessary), it is a noninvasive technique that has no potential for morbidity or complications encountered in biopsy, multiple voxel selection
1006
may be done and can be repeated during examination, and it is a single breath-hold technique using a short scan time reducing the total scan time yielding high diagnostic accuracy for liver fat quantification. REFERENCES [1] Chen CL, Fan ST, Lee SG, Makuuchi M, Tanaka K. Livingdonor liver transplantation: 12 years of experience in Asia. Transplantation 2003;75:S6e11. [2] Cheng YF, Chen CL, Lai CY, et al. Assessment of donor fatty livers for liver transplantation. Transplantation 2001;71:1221e5. [3] Chen YS, Cheng YF, De Villa VH, et al. Evaluation of living liver donors. Transplantation 2003;75:S16e9. [4] Maruzzelli L, Parr AJ, Miraglia R, Tuzzolino F, Luca A. Quantification of hepatic steatosis: a comparison of computed tomography and magnetic resonance indices in candidates for living liver donation. Acad Radiol 2014;21:507e13. [5] Son JY, Lee JY, Yi NJ, et al. Hepatic steatosis: assessment with acoustic structure quantification of US imaging. Radiology 2016;278(1):257e64. [6] Strauss S, Gavish E, Gottlieb P, Katsnelson L. Interobserver and intraobserver variability in the sonographic assessment of fatty liver. Am J Roentgenol 2007;189:W320e3. [7] Sun T, Lin X, Chen K. Evaluation of hepatic steatosis using dual-energy CT with MR comparison. Front Biosci 2014;19:1377e85. [8] Limanond P, Raman SS, Lassman C, et al. Macrovesicular hepatic steatosis in living related liver donors: correlation between CT and histologic findings. Radiology 2004;230:276e80.
CHIANG, CHANG, CHIANG ET AL [9] Peng CJ, Yuan D, Li B, et al. Body mass index evaluating donor hepatic steatosis in living donor liver transplantation. Transplant Proc 2009;41:3556e9. [10] Park SH, Kim PN, Kim KW, et al. Macrovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006;239:105e12. [11] van Werven JR, Marsman HA, Nederveen AJ, et al. Assessment of hepatic steatosis in patients undergoing liver resection: comparison of US, CT, T1-weighted dual-echo MR imaging, and point-resolved 1H MR spectroscopy. Radiology 2010;256: 159e68. [12] Lee SS, Park SH. Radiologic evaluation of nonalcoholic fatty liver disease. World J Gastroenterol 2014;20:7392e402. [13] Hines CD, Frydrychowicz A, Hamilton G, et al. T(1) independent, T(2) (*) corrected chemical shift based fat-water separation with multi-peak fat spectral modeling is an accurate and precise measure of hepatic steatosis. J Magn Reson Imaging 2011;33: 873e81. [14] Qayyum A. MR spectroscopy of the liver: principles and clinical applications. Radiographics 2009;29:1653e64. [15] Sijens PE, Edens MA, Bakker SJ, Stolk RP. MRI-determined fat content of human liver, pancreas and kidney. World J Gastroenterol 2010;16:1993e8. [16] Hamilton G, Middleton MS, Bydder M, et al. Effect of PRESS and STEAM sequences on magnetic resonance spectroscopic liver fat quantification. J Magn Reson Imaging 2009;30: 145e52. [17] Chiang HJ, Lin LH, Li CW, et al. Magnetic resonance fat quantification in living donor liver transplantation. Transplant Proc 2014;46:666e8.
Department of Diagnostic Radiology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Niao-Sung, Kaosiung, Taiwan; and the bDepartment of Surgery, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Niao-Sung, Kaosiung, Taiwan
ABSTRACT Objective. The purpose of this study is to evaluate the accuracy rate of the one breathhold single voxel hydrogen-1 magnetic resonance spectroscopy (MRS) in comparison with intraoperative biopsy for liver fat quantification in living-donor liver transplantation. Materials and Methods. A total of 80 living liver donors participated in this study. Each patient underwent both MRS and intraoperative biopsy for evaluation of liver fatty content. MRS was performed using 1.5-T magnetic resonance imaging and placed in segments 2e4, 5e8, and left lateral segment for each donor. Accuracy was assessed through receiver operating characteristic curve analysis. Sensitivity and specificity of MRS fat fractions were also calculated. Results. Eighty living-donor liver transplantation donors were enrolled in this study. There was no fatty liver in 59 subjects (73.8%), 5% to 10% fatty liver in 17 subjects, 11% to 15% fatty liver in 3 subjects, and >16% fatty liver in 1 subject. MRS fat fraction showed excellent parameters to predict between normal liver and fatty liver groups (1.85% 0.98, 8.13% 3.52, respectively; P < .0001). Linear regression between MRS fat fraction and pathology grading showed high correlation (R2 ¼ 0.7092). Pearson correlation revealed high correlation between MRS and pathology results (r ¼ 0.936), poor correlation between body mass index and pathology results (r ¼ 0.390). The sensitivity and specificity for detection of liver steatosis in MRS fat fraction were 95.2% and 98.3%, respectively. Conclusion. 1H MRS fat fraction is a highly precise and accurate method in quantification of hepatic steatosis for the living donor and can be finished in a single breath-hold.
L
IVER transplantation is the best treatment modality for end-stage liver disease [1]. Because of the shortage of deceased donors, living-donor liver transplantation (LDLT) has become a primary treatment method. The most important ethical concern about LDLT is donor safety, as it has both surgical and health risks. Hepatic steatosis quantification is critical for donor selection in LDLT because graft steatosis is associated with an increased risk of complications after liver transplantation for both donor and recipient [2]. Hepatic steatosis, which is a common finding in living liver donors, not only influences the outcome of liver transplantation for the recipient but it also affects the recovery of the living donor after partial hepatectomy [3]. Therefore, many noninvasive imaging analysis methods are
used to quantify hepatic steatosis for preoperative livingdonor evaluation [4].
This work was supported by a grant from the Ministry of Health and Welfare, Taiwan (MOHW104-TDU-B-212-124-004 to Chen CL), and health and welfare surcharge of tobacco products. and grant CMRPG8C0011 from the Chang Gung Memorial Hospital research grant, Taiwan, and a Chang Gung Medical Foundation Institutional Review Board. Taiwan approval has been obtained (101-3673B). *Address correspondence to Yu-Fan Cheng, MD, Department of Diagnostic Radiology, Kaohsiung Chang Gung Memorial Hospital, 123 Ta-Pei Road, Niao-Sung, Kaosiung 833, Taiwan. E-mail: [email protected]
ª 2016 Elsevier Inc. All rights reserved. 360 Park Avenue South, New York, NY 10010-1710
0041-1345/16 http://dx.doi.org/10.1016/j.transproceed.2015.10.068
Transplantation Proceedings, 48, 1003e1006 (2016)
1003
1004
Medical imaging has progressed rapidly in recent years. Several noninvasive methods have been developed for quantification of fat content, one of those is acoustic structure quantification of ultrasonography, which allows detection of hepatic steatosis that is 10% or greater in living liver donors [5]. This new technique has increased sensitivity in quantification of hepatic steatosis [6]; but there is still need for more precise quantification for living donor selection. Dual energy computed tomography (CT) fat analysis provides a noninvasive technique for identifying hepatic steatosis and has strongly correlated with histopathologic results [7]. However, radiation exposure has become a major problem in CT. Another limitation in CT imaging is that occasionally, pathologic abnormalities of the hepatic parenchyma are encountered such as iron overload that can influence the hepatic attenuation [8], and lead to masking or underestimation of hepatic steatosis in CT images. Magnetic resonance imaging is a highly sensitive tool for detection and characterization of fatty infiltration of liver. Magnetic resonance spectroscopy (MRS) has been used for detection and quantification of fatty infiltration in liver in recent years. MRS is one of the most accurate noninvasive techniques in assessment of hepatic steatosis. In this study, the investigators describe the 1-breath-hold stimulated echo acquisition mode (STEAM) hydrogen-1 MRS for quantification of hepatic steatosis; results are compared with liver biopsy results in the living liver donor. MATERIALS AND METHODS This prospective study was institutional review board approved by the human research committee of our institution.
Patients From March 2013 to May 2014, a total of 80 living donors (35 males and 45 females) underwent both pretransplantation MRS and intra-operative liver biopsy for liver fat quantification. Demographic characteristics and body mass index (BMI) were recorded. The mean age for all patients was 30.24 7.7 years (range 18 years to 47 years). The BMI ranged from 16.4 kg/m2 to 33.3 kg/m2 with a mean value of 23.4 kg/m2 and standard deviation of 4.1 kg/m2.
Image Acquisition Technique All the 1H MRSs were performed on a 1.5-T MR scanner (Discovery 450; GE Healthcare, Milwaukee, Wisconsin, United States). A body coil was used for signal excitation and an 8-channel body-phased array coil was used for signal reception. Breath-hold single-voxel MRS was acquired using the STEAM sequence. Voxel size was 2.0 cm3 2.0 cm3 2.0 cm3. The breath hold T1-weighted in/out phase images were used to locate the voxels placed in segments 2e4, 5e8, and left lateral segment for each donor; visible blood vessels or bile duct structures were avoided. Acquisition parameters for MRS are with a repetition time of 2,500 ms, an echo time of 12 ms, 2 next, 4 number of scan in each acquisition. In all cases, the quality of the shimming obtained in the voxel was controlled by the spectral line width (full width of half
CHIANG, CHANG, CHIANG ET AL maximum in Hz) of the unsuppressed water, obtained by the automated optimization sequence before scanning. No water suppression was applied to calculate the fat fraction of the liver. Total scan time was 21 seconds with a single breath-hold. 1
H MRS Post-processing
To estimate the liver fat fraction, all magnetic resonance spectra were analyzed with the spectral analysis program (SAGE 7.0; GE Healthcare) and performed with the same operator who has received spectrum analysis training. Post-processing steps include 8-channel signal combination, apodization, zero filling, Fourier transform, automated phase correction, and Marquardt curve fitting. The lipid signal peak was defined at 1.3 ppm, and the water was at 4.7 ppm. The signal fat fraction can then be given as the fat integral signal divided by the integral of the water and fat peaks areas.
Donor Biopsy Zero-hour biopsy specimens were obtained by wedge resection during surgery; sampling location was different for each subject, depending on the graft donated. Two independent pathologists performed histological grading of macrovesicular steatosis. For severity of fatty change and the presence of lobular inflammation, results were reported as a quantitative evaluation of the percentage of hepatocytes.
Statistical Analysis To determine the accuracy of the 1H MRS fat fraction, pathology grading was used as the gold standard. Hepatic steatosis from pathology reports in this study were divided into four groups. Group 1 included specimens normal to <5% fatty liver. Group 2 included specimens with 5% to 10% fatty liver. Group 3 included specimens with 11% to 15% fatty liver. Group 4 included specimens with >15% fatty liver. For statistical analysis, the pathology data were divided into two groups: normal (0 to <5% fatty change) and fatty liver (5% fatty change). The statistical analysis was based on independent paired Student t test. MRS fat fraction results were used to divide the subjects into two groups. Linear regression was used to detect the correlation between MRS fat fraction and pathology data. Receiver operating characteristic curve analysis was performed to analyze sensitivity and specificity for detection of steatosis in MRS fat fraction. Pearson correlation was used for the correlation among MRS fat fraction, BMI, and pathology data. All statistical analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, Illinois, United States). A P < .05 was considered significant.
RESULTS
Eighty LDLT donors were evaluated for fatty liver change by use of both MRS and intraoperative biopsy. Pathology results showed no fatty content in 73.8% (n ¼ 59), and fatty liver content in 26.2% (n ¼ 21). Of those with fatty liver, there were 21.2% with 5% to 10% fatty liver (n ¼ 17), 3.8% with 11% to 15% fatty liver (n ¼ 3), and 1.2% with >16% fatty liver (n ¼ 1). Independent paired Student t test showed that 1H MRS fat fraction was an excellent parameter to predict between normal and fatty liver groups (1.85% 0.98%, 8.13%
MAGNETIC RESONANCE SPECTROSCOPY
3.52%, P < .0001) in living donors, but no significant change in BMI (22.31 kg/m2 3.75, 26.45 kg/m2 3.65, P ¼ .863). Good linear regression was noted between 1H MRS and histopathologic results (R2 ¼ 0.7092). Pearson correlation showed a high correlation between 1H MRS and histopathologic results (r ¼ 0.936, P < .0001); however, a poor correlation is noted between BMI and histopathologic results (r ¼ 0.390, P < .0001). The sensitivity and specificity for detection of liver steatosis in 1H MRS were 95.2% and 98.3%, respectively (cutoff 4.965, area 0.999, P < .0001, 95% confidence interval 0.000e1.000) in receiver operating characteristic curve analysis. DISCUSSION
Because of the shortage of deceased donors for liver transplant, LDLT has become a widely accepted modality for treatment of end-stage liver disease in Taiwan, the top priority being donor safety [1]. Hepatic steatosis not only increases the risk of nonfunctioning or poorly functioning grafts, but it also increases the donor’s risk for postoperative complications. Regeneration of the donor’s remnant liver after major resection is impeded by hepatic steatosis. BMI has a positive correlation with increasing steatosis (r ¼ 0.390, P < .0001). Peng et al suggested that individuals with a BMI > 27.5 kg/m2 were most likely to show moderate steatosis and liver biopsy should be performed; on the other hand, donors with BMI < 23 kg/m2 were likely to show no or mild steatosis [9]. This study showed that among 17 donors with BMI > 27.5, 7 subjects (41.2%) exhibited no fatty change by biopsy. Among 43 donors with BMI < 23, 5 subjects (11.6%) demonstrated fatty change by biopsy. One donor with BMI ¼ 21.2 revealed 15% fatty change. In this regard, BMI value has no significant difference between the normal and fatty liver groups (22.31% 3.75, 26.45% 3.65, P ¼ .863). In our daily practice, the BMI was used as a reference in surveys for suitable donors in LDLT; in addition, BMI > 28 is also an indicator for liver biopsy. Unenhanced CT is being used for preoperative evaluation of fat content of donor liver. It has high diagnostic accuracy and high specificity for detecting more than 30% steatosis [10]. There is a paucity of literature to validate the utility of CT for quantitative assessment of liver steatosis and may not be as accurate for detection of mild steatosis [11,12]. Furthermore, radiation exposure makes unenhanced CT unsuitable for frequent monitoring of hepatic steatosis in potential living liver donors. In this study, MRS is used for quantification of liver fat fraction. The results showed high sensitivity and specificity in detection of steatosis, the cutoff value for detection of hepatic steatosis being 4.965% (close to 5%). MRS proves to be highly sensitive for detecting fatty liver change even in small amounts and this may be more suitable for preoperative evaluation. Biopsy is the gold standard for steatosis quantification, but it also shows sampling errors due to the heterogeneity of steatosis in the liver [13]. For this reason, three different
1005
voxels were placed to measure different segments of fat fraction to prevent sampling error and for greater precision of fat content quantification. The linear regression between 1 H MRS and histopathologic results in fatty group was 0.7092, taking into consideration the biopsy versus the MRS voxel locations which may not be the same; this variation may need more validation. The sensitivity and specificity achieved were 95.2% and 98.3%, respectively, for detection of liver steatosis in living donors. MRS allows the demonstration of relative tissue metabolite concentrations. Different metabolites have their own frequency, and the area under the signal peak in spectrum is equal to the compound concentration in the voxel [14]. Various magnetic resonance spectrum scan protocols can be used to detect fat content in the liver parenchyma. Spectroscopic imaging and chemical shift imaging use breathgating technique, which can do multi-voxel liver screen but may be time consuming (21 minutes) [15]. STEAM sequence uses a single-voxel technique to detect liver fat content. This sequence has less chance for overestimation of fat fraction due to J coupling effect in liver fat quantification and greater than point resolved spectroscopy sequence in liver fat quantification [16]. A single acquisition STEAM sequence was used in this study with a short scan time of 21 seconds and a single breath-holding technique to avoid respiratory movement problems and sampling bias during gating technique. Chemical shift imaging and iterative decomposition of water and fat with echo asymmetry and least squares estimation fat fraction imaging can obtain liver fat fraction for quantification of hepatic steatosis in the liver [17], but this is only the ratio not the fat concentration. MRS is the only noninvasive modality that can be used to determine the absolute liver fat concentration in liver. The concentration of fat in the voxel is equal to the area under the fat peak in spectrum; but the absolute quantitative value of liver fat concentration is more complex, which is not within the scope of this report because this study only discusses the fraction between the fat and water signal in spectroscopy in LDLT. 1 H MRS in vivo has limitations caused by complex spectrum analysis and technical issues. However, in this study, analysis of lipid peak (1.3 ppm) and water peak (4.7 ppm) to quantify fat fraction in the liver was performed with the use of commercialized post-processing software that can provide a semiautomatic spectrum analysis procedure. It can automatically calculate the area under the peak selected in the spectrum and instantly verify and modify the spectrum result. Therefore, the spectrum analysis has become easier to use and is faster for liver fat quantification. In conclusion, there are numerous benefits of the 1H MRS in quantification of fat infiltration of the liver for LDLT evaluation, including less systematic error, no radiation exposure (thus making it more reproducible for frequent monitoring when necessary), it is a noninvasive technique that has no potential for morbidity or complications encountered in biopsy, multiple voxel selection
1006
may be done and can be repeated during examination, and it is a single breath-hold technique using a short scan time reducing the total scan time yielding high diagnostic accuracy for liver fat quantification. REFERENCES [1] Chen CL, Fan ST, Lee SG, Makuuchi M, Tanaka K. Livingdonor liver transplantation: 12 years of experience in Asia. Transplantation 2003;75:S6e11. [2] Cheng YF, Chen CL, Lai CY, et al. Assessment of donor fatty livers for liver transplantation. Transplantation 2001;71:1221e5. [3] Chen YS, Cheng YF, De Villa VH, et al. Evaluation of living liver donors. Transplantation 2003;75:S16e9. [4] Maruzzelli L, Parr AJ, Miraglia R, Tuzzolino F, Luca A. Quantification of hepatic steatosis: a comparison of computed tomography and magnetic resonance indices in candidates for living liver donation. Acad Radiol 2014;21:507e13. [5] Son JY, Lee JY, Yi NJ, et al. Hepatic steatosis: assessment with acoustic structure quantification of US imaging. Radiology 2016;278(1):257e64. [6] Strauss S, Gavish E, Gottlieb P, Katsnelson L. Interobserver and intraobserver variability in the sonographic assessment of fatty liver. Am J Roentgenol 2007;189:W320e3. [7] Sun T, Lin X, Chen K. Evaluation of hepatic steatosis using dual-energy CT with MR comparison. Front Biosci 2014;19:1377e85. [8] Limanond P, Raman SS, Lassman C, et al. Macrovesicular hepatic steatosis in living related liver donors: correlation between CT and histologic findings. Radiology 2004;230:276e80.
CHIANG, CHANG, CHIANG ET AL [9] Peng CJ, Yuan D, Li B, et al. Body mass index evaluating donor hepatic steatosis in living donor liver transplantation. Transplant Proc 2009;41:3556e9. [10] Park SH, Kim PN, Kim KW, et al. Macrovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006;239:105e12. [11] van Werven JR, Marsman HA, Nederveen AJ, et al. Assessment of hepatic steatosis in patients undergoing liver resection: comparison of US, CT, T1-weighted dual-echo MR imaging, and point-resolved 1H MR spectroscopy. Radiology 2010;256: 159e68. [12] Lee SS, Park SH. Radiologic evaluation of nonalcoholic fatty liver disease. World J Gastroenterol 2014;20:7392e402. [13] Hines CD, Frydrychowicz A, Hamilton G, et al. T(1) independent, T(2) (*) corrected chemical shift based fat-water separation with multi-peak fat spectral modeling is an accurate and precise measure of hepatic steatosis. J Magn Reson Imaging 2011;33: 873e81. [14] Qayyum A. MR spectroscopy of the liver: principles and clinical applications. Radiographics 2009;29:1653e64. [15] Sijens PE, Edens MA, Bakker SJ, Stolk RP. MRI-determined fat content of human liver, pancreas and kidney. World J Gastroenterol 2010;16:1993e8. [16] Hamilton G, Middleton MS, Bydder M, et al. Effect of PRESS and STEAM sequences on magnetic resonance spectroscopic liver fat quantification. J Magn Reson Imaging 2009;30: 145e52. [17] Chiang HJ, Lin LH, Li CW, et al. Magnetic resonance fat quantification in living donor liver transplantation. Transplant Proc 2014;46:666e8.