Evaluation of liver function using gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid enhanced magnetic resonance imaging based on a three-dimensional volumetric analysis system

Evaluation of liver function using gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid... Background Magnetic resonance imaging with gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (EOB-MRI) is a diagnostic modality for liver tumors. Three-dimensional (3D) volumetric analysis systems using EOB-MRI data are used to simulate liver anatomy for surgery. This study was conducted to investigate clinical utility of a 3D volumetric analysis system on EOB-MRI to evaluate liver function. Methods Between August 2014 and December 2015, 181 patients underwent laboratory and radiological exams as standardized preoperative evaluation for liver surgery. The liver-spleen contrast-enhanced ratio (LSR) was measured by a semi-automated 3D volumetric analysis system on EOB-MRI. First, the inter-evaluator variability of the calculated LSR was evaluated. Additionally, a subset of liver surgical specimens was evaluated histologically by using immunohisto- chemical staining. Finally, the correlations between the LSR and grading systems of liver function, laboratory data, or histological findings were analyzed. Results The inter-evaluator correlation coefficient of the measured LSR was 0.986. The mean LSR was significantly correlated with the Child–Pugh score (p = 0.014) and the ALBI score (p \ 0.001). Significant correlations were also observed between the LSR and indocyanine green retention rate at 15 min (r = - 0.601, p \ 0.001), between the LSR and liver fibrosis stage (r = - 0.556, p \ 0.001), and between the LSR and liver steatosis grade (r = - 0.396, p \ 0.001). Conclusion The LSR calculated by a 3D volumetric analysis system on EOB-MRI was highly reproducible and was shown to be correlated with liver function parameters and liver histology. These data suggest that this imaging modality can be a reliable tool to evaluate liver function. Keywords Magnetic resonance imaging  Gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid Three-dimensional volumetric analysis system  Liver function  Liver steatosis Introduction After major hepatectomy, the reported incidence of liver Electronic supplementary material The online version of this failure is 3–8%, and the reported rate of mortality associ- article (https://doi.org/10.1007/s12072-018-9874-x) contains ated with liver failure is approximately 5% [1]. supplementary material, which is available to authorized users. Department of Pathology and Clinical Laboratory, National & Naoto Gotohda Cancer Center Hospital East, Kashiwa, Japan ngotohda@east.ncc.go.jp Course of Advanced Clinical Research of Cancer, Juntendo 1 University Graduate School of Medicine, Tokyo, Japan Department of Hepatobiliary and Pancreatic Surgery, National Cancer Center Hospital East, 6-5-1 Kashiwa-no-ha, Kashiwa, Chiba 277-8577, Japan Department of Diagnostic Radiology, National Cancer Center Hospital East, Kashiwa, Japan 123 Hepatology International (2018) 12:368–376 369 Preoperative precise assessment of liver function, extent of were used to analyze the correlations between the LSR and resection, and estimated remnant liver volume is important grading systems of liver function or laboratory data (patient to minimize the risks associated with liver surgery. cohort 2). Finally, a subgroup of all 112 patients who Gadolinium ethoxybenzyl diethylenetriamine pen- underwent liver resection was used to analyze the corre- taacetic acid (Gd-EOB-DTPA) is a liver-specific magnetic lations between the LSR and histological parameters (pa- resonance imaging (MRI) contrast agent. MRI using Gd- tient cohort 3). The patient flow chart of this study is shown EOB-DTPA (EOB-MRI) is widely used to detect and in Fig. 1. characterize hepatocellular carcinoma or metastatic liver tumors [2]. Gd-EOB-DTPA has a well-known metabolic MRI protocol pathway, and several studies have suggested that EOB- MRI is useful to evaluate liver function [3, 4]. The liver-to- All MRI studies were performed using one of the two 3.0 T spleen signal intensity ratio (LSR) on EOB-MRI has been scanners at our institution (Achieva or Ingenia, Philips used as a parameter to assess liver function [5–8], but the Medical Systems, Amsterdam, Netherlands). Contrast-en- conventional method to measure the LSR using two-di- hanced 3D fat-suppressed T1-weighted images were mensional (2D) regions of interest might be affected by obtained 20 min after intravenous administration of Gd- sampling errors or inter-evaluator variability [9]. EOB-DTPA for hepatobiliary phase imaging, using the In a related field, three-dimensional (3D) volumetric following parameters: repetition time (TR) 4 ms, echo time analysis systems are being used to simulate liver anatomy (TE) 2 ms, flip angle 10, slice thickness 4.6 mm, matrix for surgery. Using these systems, inter-evaluator variability size 512 9 512 (Achieva), and TR 3 ms, TE 2 ms, flip is expected to be reduced, as signal intensity from the angle 10, slice thickness 4.6 mm, matrix size 480 9 480 whole liver can be included in a semi-automatic analysis. (Ingenia). Gd-EOB-DTPA was administered at a dose of There have been no previous clinical studies in which the about 0.1 mL/kg, by rapid intravenous bolus injection 3D volumetric analysis system was used to measure the using a power injector (SONIC SHOT GX, NEMOTO- LSR on EOB-MRI. Therefore, the objective of this study KYORINDO, Tokyo, Japan), at a rate of 2 mL/s. was to evaluate the variability of the calculated LSR of EOB-MRI using a 3D volumetric analysis system, and to MRI data analysis investigate the correlations between the LSR and liver function parameters or histological findings. The LSR was calculated using images from the 20 min- delayed hepatobiliary phase, with the 3D volumetric analysis system SYNAPSE VINCENT (Fujifilm Medical, Materials and methods Tokyo, Japan), by a single investigator (Ma.Ku.), under supervision of an experienced radiologist (T.K.). First, the Patients investigator placed small operator-defined volumes of interest (VOIs), one in the liver and another in the spleen Between August 2014 and December 2015, a total of 304 parenchyma, avoiding vessels and tumors (Fig. 2a). Sec- consecutive patients underwent laboratory and radiological ond, the liver and spleen parenchyma were semi-automat- examinations as preoperative evaluations in consideration ically extracted using the image-processing algorithm of liver surgery at the National Cancer Center Hospital (Fig. 2b). Finally, the LSR was calculated as the average East, Japan. Of the 304 patients, 123 were excluded for the liver parenchyma signal intensity (Fig. 2c) divided by the following reasons: contraindications to EOB-MRI average spleen parenchyma signal intensity. Using the (n = 29), inconsistent MRI acquisition technique (n = 52), patient cohort 1, the variability of the calculated LSR incomplete laboratory data (n = 3), or prior splenectomy among four evaluators (surgeons with 7–11 years of clin- (n = 1). Thirty-eight patients who had undergone portal ical experience) was analyzed. Then, since inclusion of vein embolization were also excluded because the degree extrahepatic parenchymal tissue, such as portal and hepatic of liver enhancement from Gd-EOB-DTPA could depend veins, intrahepatic bile ducts, cysts, and tumors, might considerably on the portal vein flow. affect the average liver signal intensity, these structures The remaining 181 patients who underwent EOB-MRI were subtracted manually by the investigator (Fig. 2e). of the liver using a standardized imaging technique were Additionally, the vascular subtraction LSR (vsLSR) was used for the analysis in this study. Among them, 24 con- calculated, and its correlation with the LSR was analyzed secutive patients who underwent preoperative evaluation using cohort 1 (Fig. 1). from November 2015 to December 2015 were used to examine the variability of the LSR among four different evaluators (patient cohort 1). Then, all the 181 patients 123 370 Hepatology International (2018) 12:368–376 Fig. 1 Flow-diagram describing the patient cohorts in this study. Gd-EOB-DTPA gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid; LSR liver-to-spleen signal intensity ratio Grading systems of liver function and laboratory stained by hematoxylin and eosin (HE), azo carmine ani- line blue (AZAN), and smooth muscle actin (SMA) data (Fig. 3). The specimen slides were scanned using the The following laboratory parameters were obtained preop- NanoZoomer 2.0 system (Hamamatsu Photonics, Hama- matsu, Japan), and morphometric analysis was performed eratively and within a month before or after EOB-MRI were collected from the patients’ medical records: indocyanine using the WinROOF 6.5 image processing software pro- gram (MITANI Corporation, Tokyo, Japan). Histological green retention rate at 15 min (ICGR15), white blood cell count, platelet count, prothrombin activity, and serum levels analyses were performed by a single investigator (Ma.Ku.), under supervision of an experienced pathologist (Mo.Ko.). of hemoglobin, albumin, total bilirubin, aspartate amino- transferase, alanine aminotransferase, and creatinine. Grad- Liver fibrosis was assessed in three ways. First, using HE-stained slides, the fibrosis stage was morphologically ing systems of liver function, such as the Child–Pugh score and the albumin-bilirubin (ALBI) score, were analyzed for categorized by the METAVIR scoring system as follows: F0, no fibrosis; F1, portal fibrosis without septa; F2, few the correlation with the LSR. The Child–Pugh score was graded into A, B, or C, using five variables, including septa; F3, numerous septa without cirrhosis; F4, cirrhosis bilirubin, albumin, prothrombin, ascites status, and degree of [10]. Second, the ratio of fibrosis (ROF) on AZAN-stained slides (ROFazan) was quantified using morphometric encephalopathy. The ALBI score was calculated as follows: analysis from a color-detecting algorithm (Fig. 3b), using 0:085 ðalbumin g/lÞþ 0:66 the same technique as reported in a previous study [11]. logðtotal bilirubin lmol/lÞ; Third, the ROF on SMA-stained slides (ROFsma) was and categorized into the following three grades: quantified in a similar manner (Fig. 3d). Liver steatosis was also assessed in two ways. First, the \  2:60; ALBI 1; [ 2:06 to  1:39; ALBI 2; steatosis grade was morphologically categorized as in [  1:39; ALBI 3: Kleiner et al. [12]: grade 0, \ 5%; grade 1, 5–33%; grade 2, 34–66%; grade 3, [ 67%. Second, the ratio of droplet Using patient cohort 2, the correlations between the LSR steatosis (ROS) was quantified using morphometric anal- and grading systems of liver function or other laboratory ysis from the color-detecting algorithm (Fig. 3f). data were analyzed (Fig. 1). Statistical analysis Histological analysis of the surgical specimen Statistical analyses were performed using JMP (version From the 181 studied patients, 112 patients underwent liver 12.0.10; SAS Institute, Cary, NC). Ebel’s intraclass cor- resection (patient cohort 3) (Fig. 1). Their surgical speci- relation coefficients were used to evaluate the inter-eval- mens of non-tumoral liver tissues were fixed in formalin uator variability of the LSR and the correlation between the and embedded in paraffin. Then, 4-lm-thick sections were 123 Hepatology International (2018) 12:368–376 371 Fig. 2 Image analysis using the three-dimensional volumetric system. calculated semi-automatically. d Liver parenchyma extracted before a The investigator placed a small volume of interest (VOI) in the liver subtracting vascular structures. e Liver parenchyma after subtracting parenchyma. b The liver parenchyma was semi-automatically vascular structures, such as portal veins and hepatic veins, and extracted from the initial VOI seed. c The three-dimensional liver intrahepatic bile ducts volume was extracted, and the average liver signal intensity was LSR and vsLSR in patient cohort 1. Correlations between hepatocellular carcinoma in 94 patients (50%), metastatic the LSR and the clinicopathological factors were assessed liver cancer in 76 patients (42%), perihilar cholangiocar- by the standard Pearson’s correlation coefficient or cinoma in five patients (3%), cholangiocellular carcinoma Spearman’s rank correlation coefficient in patient cohorts 2 in four patients (2%), and other disease in six patients and 3. The correlations between the LSR and fibrosis stage (3%). The primary lesion of metastatic liver cancer was or steatosis grade were evaluated using pairwise compar- colon in 71 patients, stomach in two patients, pancreas in isons with the Mann–Whitney test in patient cohort 3. Two- one patient, lung in one patient, and biliary tract in one sided p values of less than 0.05 were considered indicative patient, respectively. There was an underlying liver infec- of significance. tion of the hepatitis C virus in 44 patients (24%) and of the hepatitis B virus in nine patients (5%). Of the 181 patients undergoing EOB-MRI of the liver in this study, 177 Results (97.8%) patients were classified as having the Child–Pugh grade A and four (2.2%) patients as having the Child–Pugh Patients’ demographics grade B. The median value of ICGR15 was 13.1% (range, 2.9–58.9%). The therapies provided for the liver tumors The median age was 70 (range, 39–90), and 130 (72%) of were as follows: surgical resection in 112 patients (62%), 181 patients were male. The diagnoses were as follows: chemotherapy in 26 patients (14%), radiofrequency 123 372 Hepatology International (2018) 12:368–376 Fig. 3 Areas of fibrosis and steatosis were calculated using morphometric analysis of color- detecting algorithm (WinROOF software, version 6.5; MITANI Corporation, Tokyo, Japan). a Azo carmine aniline blue (AZAN) and c smooth muscle actin (SMA) stain. The AZAN/ SMA-positive area was determined using a color- detecting algorithm and is represented as bright green in b and d. The ratio of fibrosis (ROF) was calculated as the percentage area of the entire field and the AZAN/SMA- positive area. Hematoxylin– eosin stain of fatty liver (e). Fat droplets were determined using a color-detecting algorithm and are represented as bright green in image (f). The ratio of steatosis (ROS) was calculated as the percentage area of the entire field and the fat droplets area ablation in 11 patients (6%), transcatheter arterial Correlation between the LSR and grading chemoembolization in 13 patients (7%), proton beam systems of liver function (patient cohort 2, radiation therapy in 12 patients (7%), and supportive care n = 181) in seven patients (4%). Correlations between the LSR and grading systems of liver Inter-evaluator variability of the LSR function were summarized in Table 1. The mean LSR was and correlation between the LSR and vsLSR lower in the patients with the Child–Pugh grade B than (patient cohort 1, n = 24) those with the Child–Pugh grade A (1.57 vs 1.97, p = 0.014). The mean LSR was lower in the patients with The intraclass correlation coefficient of the LSR amongst the ALBI score 2 than those with the ALBI score 1 (1.64 vs the four evaluators calculated using the 3D volumetric 2.05, p \ 0.001). analysis system, was 0.986. The intraclass correlation coefficient between the LSR and vsLSR, evaluated by a single investigator, was 0.987. 123 Hepatology International (2018) 12:368–376 373 Table 1 Correlations between the liver-to-spleen ratio (LSR) and aminotransferase (r = - 0.422, p \ 0.001), and alanine grading systems of liver function (patient cohort 2, n = 181) aminotransferase (r = - 0.287, p \ 0.001). n = 181 LSR (mean) p Correlations between the LSR and histological Child–Pugh score findings (patient cohort 3, n = 112) A 177 (97.8%) 1.97 = 0.014 B 4 (2.2%) 1.57 Correlations between the LSR and histological findings, ALBI score including liver fibrosis and steatosis, are summarized in 1 142 (78.5%) 2.05 \ 0.001 Table 2 and Fig. 4. In terms of liver fibrosis, negative 2 39 (21.5%) 1.64 correlations were observed between the LSR and the METAVIR score (r = - 0.556, p \ 0.001), between the LSR and ROFazan (r = - 0.424, p \ 0.001), and between the LSR and ROFsma (r = - 0.592, p \ 0.001) (Table 2). Correlations between the LSR and laboratory The LSR value was significantly greater for fibrosis stages data (patient cohort 2, n = 181) F0 and F1 than for stages F2, F3, or F4 (each pairwise comparison, p \ 0.001) (Fig. 4a). In terms of liver Correlations between the LSR and laboratory data are steatosis, negative correlations were observed between the summarized in Table 2. Positive correlations were LSR and the Kleiner’s grade (r = - 0.396, p \ 0.001) and observed between the LSR and the following parameters: between the LSR and ROS (r = - 0.428, p \ 0.001). The platelet count (r = 0.307, p \ 0.001), serum level of LSR value was significantly greater for steatosis grade 0 albumin (r = 0.453, p \ 0.001), and prothrombin activity than for grade 1 (p \ 0.001), grade 2 (p \ 0.001), or grade (r = 0.426, p \ 0.001). Negative correlations were 3 (each pairwise comparison, p = 0.001) (Fig. 4b). observed between the LSR and the following parameters: ICGR15 (r = - 0.601, p \ 0.001) and serum levels of total bilirubin (r = - 0.370, p \ 0.001), aspartate Discussion Table 2 Correlations between the liver-to-spleen ratio (LSR) and In this study, the clinical utility of the LSR, calculated as laboratory data or histological findings the signal intensity of the liver parenchyma divided by that rp of the spleen on EOB-MRI, was evaluated using a 3D volumetric analysis system. The reproducibility of the Laboratory data (patient cohort 2, n = 181) calculated LSR amongst the different evaluators was very ICGR15 - 0.601 \ 0.001 high. The LSR was shown to be correlated with grading White blood cell 0.067 0.368 systems of liver function, such as the Child–Pugh score and Platelet count 0.307 \ 0.001 the ALBI score. In addition, significant correlations Prothrombin activity 0.426 \ 0.001 between the LSR and ICGR15 and between the LSR and Hemoglobin 0.021 0.783 histological findings of liver fibrosis or steatosis were Albumin 0.453 \ 0.001 observed. This is the first report using a 3D volumetric Total bilirubin - 0.370 \ 0.001 analysis system of EOB-MRI to evaluate liver function. Aspartate aminotransferase - 0.422 \ 0.001 Conventionally, signal intensity of the liver parenchyma Alanine aminotransferase - 0.287 \ 0.001 on EOB-MRI has been measured by ROIs arbitrarily Creatinine - 0.116 0.120 selected on a few 2D MRI slices. However, such systems Histological findings (patient cohort 3, n = 112) might include a certain degree of selection bias of the Liver fibrosis ROIs. On the other hand, 3D volumetric analysis systems, METAVIR score - 0.556 \ 0.001 which extract the entire volume of specific internal organs ROF (AZAN stain) - 0.424 \ 0.001 based on high-precision image processing algorithm, have ROF (SMA stain) - 0.592 \ 0.001 been introduced in clinical settings in recent years. The Liver steatosis benefits of preoperative simulation of liver anatomy with Kleiner grade - 0.396 \ 0.001 the use of 3D image visualization technologies have been ROS - 0.428 \ 0.001 shown in several papers [13, 14]. Takamoto et al. showed a high correlation between estimated liver volume on pre- ICGR15 indocyanine green retention rate at 15 min; ROF ratio of fibrosis; ROS ratio of steatosis; AZAN azo carmine aniline blue; SMA operative computed tomography and the weight of the smooth muscle actin resected specimens [15]. Ogawa et al. reported the utility of a 3D volumetric analysis system on computed tomography 123 374 Hepatology International (2018) 12:368–376 Fig. 4 Box plots of the liver-to-spleen ratio (LSR) according to the fibrosis stage (a) and steatosis grade (b). *Significant differences in pairwise comparisons using the Mann–Whitney test images to predict the area of the liver that is embolized the strongest correlation between the LSR and ICGR15, after transcatheter arterial chemoembolization [16]. The compared to the other laboratory data analyzed. advantage of 3D volumetric analysis systems compared to There are several studies showing a significant correla- 2D ones is that they enable the calculation of the average tion between signal intensity of the liver on 2D analysis whole liver signal intensity automatically, from only a systems on EOB-MRI and liver fibrosis [6, 20–22]. These small VOI drawn in the liver parenchyma. The present studies evaluated liver fibrosis using either the METAVIR study showed that the inter-evaluator correlation coeffi- score or the New INUYAMA classification as categorical cient of the LSR values was 0.986. This means that the variables assessed by pathologists. In this study, morpho- LSR, calculated using a 3D volumetric analysis system of metric analysis, using image processing software, enabled EOB-MRI, is an objective, precise, and quantitative index the quantification of fibrosis or steatosis in the liver, and to measure liver signal intensity. showed a significantly inverse correlation between the LSR There are several studies that used 2D analysis systems and fibrosis or steatosis. These data suggest that infiltration on EOB-MRI to evaluate liver function. These studies in the liver parenchyma of stromal tissues, such as fibrosis indicated that liver signal intensity, measured from ROIs, or steatosis, decreases the relative area of hepatocytes and was correlated with liver functional reserve markers, such leads to the reduction of signal intensity of the whole liver as ICGR15, prothrombin activity, or liver fibrosis on EOB-MRI. Interestingly, as shown in the Supplemen- [3–5, 17]. The present study showed that the LSR calcu- tary Table, ICGR15 was not significantly correlated with lated using a 3D volumetric analysis system on EOB-MRI fibrosis or steatosis in the liver. These results suggest that was correlated with grading systems of liver function and the function of the liver that was affected by fibrosis or laboratory data, such as platelet count, serum levels of steatosis was not reflected accurately by ICGR15 but by the albumin, total bilirubin, aspartate aminotransferase, alanine signal intensity of the whole liver on EOB-MRI. Therefore, aminotransferase, prothrombin activity, and ICGR15. the LSR, calculated by a 3D volumetric analysis system on ICGR15 is one of the well-known biochemical indices used EOB-MRI, may indicate liver function more accurately to predict post-hepatectomy remnant liver function. In than the other liver function parameters. There might be principle, ICG is taken up to hepatocytes via the organic concerns regarding other components in the liver than anion transporting polypeptide or Na -taurocholate co- normal parenchyma, fibrosis, and steatosis. Majority of transporting polypeptides located in the sinusoidal mem- previous studies reported no significant correlations brane, and is excreted into the biliary system via the ATP- between fibrosis or steatosis and iron deposition dependent export pump multidrug resistance-associated histopathologically or radiographically [23–25]. Moreover, protein 2 without biotransformation [18]. Gd-EOB-DTPA there were no patients with hemosiderosis or hemochro- is also taken up to hepatocytes via the organic anion matosis clinically in our patient cohort. Therefore, liver transporting polypeptides, and is excreted into the biliary fibrosis or steatosis was considered to be evaluated using system via the multidrug resistance-associated protein 2 EOB-MRI regardless of the iron deposition. [19]. Therefore, in the light of the metabolic mechanism In this study, there were five patients with perihilar theory, the signal intensity of the liver on EOB-MRI would cholangiocarcinoma who developed obstructive jaundice at be expected to reflect ICGR15. The present study showed presentation. However, we consider that the influence of 123 Hepatology International (2018) 12:368–376 375 obstructive jaundice on evaluating liver function using In conclusion, the LSR calculated using a 3D volumetric EOB-MRI was expected to be minimal, because all the analysis system on EOB-MRI were highly reproducible, patients who had obstructive jaundice were re-examined by and were correlated with grading systems of liver function, EOB-MRI for operative indication after biliary drainage laboratory data, and histological findings. EOB-MRI using and resolution of jaundice. However, evaluation of liver a 3D volumetric analysis system may be a reliable modality function using EOB-MRI for the patients with obstructive to evaluate liver function. jaundice may be a subject of future study. Acknowledgements The authors would like to thank Ryo Morisue, Our study has several limitations. First, there was a Toshiyuki Suzuki and Hidetoshi Aizawa for assisting with data col- considerable number of patients (52/304, 17%) that had to lection, and Kenichi Ikejima M.D. (Department of Gastroenterology, be excluded due to a different EOB-MRI acquisition pro- Juntendo University Graduate School of Medicine, Tokyo, Japan) for tocol (mostly due to a different slice thickness). During the his technical support. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit study period, the slice thickness of 4.6 mm might be rel- sectors. atively thick for 3D volumetric analysis. The MRI protocol was determined to diagnose liver tumors but not to evaluate Compliance with ethical standards liver function in this study. However, our study showed that the LSR was correlated with the Child–Pugh grade, the Conflict of interest Masashi Kudo, Naoto Gotohda, Motokazu Sugi- ALBI score, ICGR15 and other laboratory data, and his- moto, Tatsushi Kobayashi, Motohiro Kojima, Shinichiro Takahashi, Masaru Konishi, and Ryuichi Hayashi declare that they have no tological findings, although the MRI conditions were set to conflicts of interest. diagnose liver tumors. These results suggest that EOB-MRI would be applied for evaluating liver function in the clin- Ethical approval This study protocol conformed to the ethical ical situation. Furthermore, the MRI protocol in the pre- guidelines of the 1975 Declaration of Helsinki and was approved by vious studies that evaluated liver function using EOB-MRI the institutional review board of the National Cancer Center, Japan. did not differ widely from that in our study: the slice Informed consent Informed consent was obtained from each patient thickness was 5.0 mm in the study by Okada et al. [8], included in the study. 4.0 mm by Nishie et al. [7], and 3.8 mm by Matsushima et al. [4]. Second, in the 3D volumetric analysis, extra- Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creative hepatic-parenchymal tissues such as portal vein or hepatic commons.org/licenses/by/4.0/), which permits unrestricted use, dis- vein were included in the whole liver signal intensity cal- tribution, and reproduction in any medium, provided you give culation: the LSR was the metric primarily used. However, appropriate credit to the original author(s) and the source, provide a since the study showed a high correlation between the LSR link to the Creative Commons license, and indicate if changes were made. and vsLSR (LSR excluding extrahepatic-parenchymal tis- sues), and demonstrated that the LSR adequately reflects contrast enhancement of the liver parenchyma, this sug- gests that liver function can be evaluated without sub- References tracting vessels and vascular perfusion areas. Third, two different MR scanners were used in this study. This might 1. Yadav K, Shrikhande S, Goel M. Post hepatectomy liver failure: have caused a systemic bias in the calculation of liver or concept of management. J Gastrointest Cancer. 2014;45(4):405–413. https://doi.org/10.1007/s12029-014-9646-3 spleen signal intensities. However, the difference is not Epub 2014/08/12. assumed to be significant, as the MRI protocols were 2. Huppertz A, Balzer T, Blakeborough A, Breuer J, Giovagnoni A, similar between the two different machines and all images Heinz-Peer G, et al. Improved detection of focal liver lesions at were taken by T-1 weighted technique. In a previous MR imaging: multicenter comparison of gadoxetic acid-enhanced MR images with intraoperative findings. Radiology. multicenter study, Okada et al. reported no significant 2004;230(1):266–275. https://doi.org/10.1148/radiol.2301020269 differences in signal intensity of the liver on EOB-MRI Epub 2003/12/26. between different MRI scanners [8]. Fourth, there were a 3. Utsunomiya T, Shimada M, Hanaoka J, Kanamoto M, Ikemoto T, large number of patients excluded for having prior history Morine Y, et al. Possible utility of MRI using Gd-EOB-DTPA for estimating liver functional reserve. J Gastroenterol. of portal vein embolization. Previous studies showed a 2012;47(4):470–476. https://doi.org/10.1007/s00535-011-0513-8 significant decrease of signal intensity of the liver lobe Epub 2011/12/16. after portal vein embolization [26]. Therefore, in the pre- 4. Matsushima S, Sato Y, Yamaura H, Kato M, Kinosada Y, Era S, sent study, we decided to investigate the entire liver et al. Visualization of liver uptake function using the uptake contrast-enhanced ratio in hepatobiliary phase imaging. Magn function, instead of regional liver function. Preoperative Reson Imaging. 2014;32(6):654–659. https://doi.org/10.1016/j. evaluation of the regional liver function using EOB-MRI mri.2014.02.017 Epub 2014/03/19. for major hepatectomy would be the subject of future study. 123 376 Hepatology International (2018) 12:368–376 5. Motosugi U, Ichikawa T, Sou H, Sano K, Tominaga L, Kitamura 16. Ogawa C, Minami Y, Morita M, Noda T, Arasawa S, Izuta M, T, et al. Liver parenchymal enhancement of hepatocyte-phase et al. Prediction of embolization area after conventional tran- images in Gd-EOB-DTPA-enhanced MR imaging: which bio- scatheter arterial chemoembolization for hepatocellular carci- logical markers of the liver function affect the enhancement? noma using SYNAPSE VINCENT. Dig Dis. J Magn Reson Imaging. 2009;30(5):1042–1046. https://doi.org/ 2016;34(6):696–701. https://doi.org/10.1159/000448859 Epub 10.1002/jmri.21956 Epub 2009/10/27. 2016/10/18. 6. Motosugi U, Ichikawa T, Oguri M, Sano K, Sou H, Muhi A, et al. 17. Yamada A, Hara T, Li F, Fujinaga Y, Ueda K, Kadoya M, et al. Staging liver fibrosis by using liver-enhancement ratio of Quantitative evaluation of liver function with use of gadoxetate gadoxetic acid-enhanced MR imaging: comparison with aspartate disodium-enhanced MR imaging. Radiology. aminotransferase-to-platelet ratio index. Magn Reson Imaging. 2011;260(3):727–733. https://doi.org/10.1148/radiol.11100586 2011;29(8):1047–1052. https://doi.org/10.1016/j.mri.2011.05.007 Epub 2011/06/30. Epub 2011/07/22. 18. Hoekstra LT, de Graaf W, Nibourg GA, Heger M, Bennink RJ, 7. Nishie A, Ushijima Y, Tajima T, Asayama Y, Ishigami K, Stieger B, et al. Physiological and biochemical basis of clinical Kakihara D, et al. Quantitative analysis of liver function using liver function tests: a review. Ann Surg. 2013;257(1):27–36. superparamagnetic iron oxide- and Gd-EOB-DTPA-enhanced https://doi.org/10.1097/sla.0b013e31825d5d47 Epub 2012/07/28. MRI: comparison with technetium-99 m galactosyl serum albu- 19. Van Beers BE, Pastor CM, Hussain HK. Primovist, eovist: what min scintigraphy. Eur J Radiol. 2012;81(6):1100–1104. https:// to expect? J Hepatol. 2012;57(2):421–429. https://doi.org/10. doi.org/10.1016/j.ejrad.2011.02.053 Epub 2011/03/26. 1016/j.jhep.2012.01.031 Epub 2012/04/17. 8. Okada M, Murakami T, Kuwatsuru R, Nakamura Y, Isoda H, 20. Watanabe H, Kanematsu M, Goshima S, Kondo H, Onozuka M, Goshima S, et al. Biochemical and clinical predictive approach Moriyama N, et al. Staging hepatic fibrosis: comparison of and time point analysis of hepatobiliary phase liver enhancement gadoxetate disodium-enhanced and diffusion-weighted MR on Gd-EOB-DTPA-enhanced MR images: a multicenter study. imaging—preliminary observations. Radiology. Radiology. 2016;281(2):474–483. https://doi.org/10.1148/radiol. 2011;259(1):142–150. https://doi.org/10.1148/radiol.10100621 2016151061 Epub 2016/05/20. Epub 2011/01/21. 9. Imura S, Shimada M, Utsunomiya T. Recent advances in esti- 21. Nojiri S, Kusakabe A, Fujiwara K, Shinkai N, Matsuura K, Iio E, mating hepatic functional reserve in patients with chronic liver et al. Noninvasive evaluation of hepatic fibrosis in hepatitis C damage. Hepatol Res. 2015;45(1):10–19. https://doi.org/10.1111/ virus-infected patients using ethoxybenzyl-magnetic resonance hepr.12325 Epub 2014/03/13. imaging. J Gastroenterol Hepatol. 2013;28(6):1032–1039. https:// 10. Bedossa P, Poynard T. An algorithm for the grading of activity in doi.org/10.1111/jgh.12181 Epub 2013/02/26. chronic hepatitis C. The METAVIR cooperative study group. 22. Shimamoto D, Nishie A, Asayama Y, Ushijima Y, Takayama Y, Hepatology. 1996;24(2):289–293. https://doi.org/10.1002/hep. Fujita N, et al. MR prediction of liver function and pathology 510240201 Epub 1996/08/01. using Gd-EOB-DTPA: effect of liver volume consideration. 11. Sugimoto M, Takahashi S, Kojima M, Kobayashi T, Gotohda N, Biomed Res Int. 2015;2015:141853. https://doi.org/10.1155/ Konishi M. In patients with a soft pancreas, a thick parenchyma, a 2015/141853 Epub 2015/11/27. small duct, and fatty infiltration are significant risks for pancreatic 23. Franca M, Alberich-Bayarri A, Marti-Bonmati L, Oliveira P, fistula after pancreaticoduodenectomy. J Gastrointest Surg. 2017. Costa FE, Porto G, et al. Accurate simultaneous quantification of https://doi.org/10.1007/s11605-017-3356-7. liver steatosis and iron overload in diffuse liver diseases with 12. Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, MRI. Abdom radiol (New York). 2017;42(5):1434–1443. https:// Cummings OW, et al. Design and validation of a histological doi.org/10.1007/s00261-017-1048-0 Epub 2017/01/23. scoring system for nonalcoholic fatty liver disease. Hepatology. 24. Szurowska E, Sikorska K, Izycka-Swieszewska E, Nowicki T, 2005;41(6):1313–1321. https://doi.org/10.1002/hep.20701 Epub Romanowski T, Bielawski KP, et al. The role of MR imaging in 2005/05/26. detection of hepatic iron overload in patients with cirrhosis of 13. Ohshima S. Volume analyzer SYNAPSE VINCENT for liver different origins. BMC Gastroenterol. 2010;10:13. https://doi.org/ analysis. J Hepatobiliary Pancreat Sci. 2014;21(4):235–238. 10.1186/1471-230x-10-13 Epub 2010/01/29. https://doi.org/10.1002/jhbp.81 Epub 2014/02/13. 25. Ludwig J, Batts KP, Moyer TP, Baldus WP, Fairbanks VF. Liver 14. Mochizuki K, Takatsuki M, Soyama A, Hidaka M, Obatake M, biopsy diagnosis of homozygous hemochromatosis: a diagnostic Eguchi S. The usefulness of a high-speed 3D-image analysis algorithm. Mayo Clin Proc. 1993;68(3):263–267 Epub 1993/03/ system in pediatric living donor liver transplantation. Ann 01. Transplant. 2012;17(1):31–34 Epub 2012/04/03. 26. Geisel D, Raabe P, Ludemann L, Malinowski M, Stockmann M, 15. Takamoto T, Hashimoto T, Ogata S, Inoue K, Maruyama Y, Seehofer D, et al. Gd-EOB-DTPA-enhanced MRI for monitoring Miyazaki A, et al. Planning of anatomical liver segmentectomy future liver remnant function after portal vein embolization and and subsegmentectomy with 3-dimensional simulation software. extended hemihepatectomy: a prospective trial. Eur Radiol. Am J Surg. 2013;206(4):530–538. https://doi.org/10.1016/j.amj 2016;27(7):3080–3087. https://doi.org/10.1007/s00330-016- surg.2013.01.041 Epub 2013/07/03. 4674-y Epub 2016/12/15. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Hepatology International Springer Journals

Evaluation of liver function using gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid enhanced magnetic resonance imaging based on a three-dimensional volumetric analysis system

Free
9 pages

Loading next page...
 
/lp/springer_journal/evaluation-of-liver-function-using-gadolinium-ethoxybenzyl-b65LEawq1e
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Hepatology; Colorectal Surgery; Surgery
ISSN
1936-0533
eISSN
1936-0541
D.O.I.
10.1007/s12072-018-9874-x
Publisher site
See Article on Publisher Site

Abstract

Background Magnetic resonance imaging with gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (EOB-MRI) is a diagnostic modality for liver tumors. Three-dimensional (3D) volumetric analysis systems using EOB-MRI data are used to simulate liver anatomy for surgery. This study was conducted to investigate clinical utility of a 3D volumetric analysis system on EOB-MRI to evaluate liver function. Methods Between August 2014 and December 2015, 181 patients underwent laboratory and radiological exams as standardized preoperative evaluation for liver surgery. The liver-spleen contrast-enhanced ratio (LSR) was measured by a semi-automated 3D volumetric analysis system on EOB-MRI. First, the inter-evaluator variability of the calculated LSR was evaluated. Additionally, a subset of liver surgical specimens was evaluated histologically by using immunohisto- chemical staining. Finally, the correlations between the LSR and grading systems of liver function, laboratory data, or histological findings were analyzed. Results The inter-evaluator correlation coefficient of the measured LSR was 0.986. The mean LSR was significantly correlated with the Child–Pugh score (p = 0.014) and the ALBI score (p \ 0.001). Significant correlations were also observed between the LSR and indocyanine green retention rate at 15 min (r = - 0.601, p \ 0.001), between the LSR and liver fibrosis stage (r = - 0.556, p \ 0.001), and between the LSR and liver steatosis grade (r = - 0.396, p \ 0.001). Conclusion The LSR calculated by a 3D volumetric analysis system on EOB-MRI was highly reproducible and was shown to be correlated with liver function parameters and liver histology. These data suggest that this imaging modality can be a reliable tool to evaluate liver function. Keywords Magnetic resonance imaging  Gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid Three-dimensional volumetric analysis system  Liver function  Liver steatosis Introduction After major hepatectomy, the reported incidence of liver Electronic supplementary material The online version of this failure is 3–8%, and the reported rate of mortality associ- article (https://doi.org/10.1007/s12072-018-9874-x) contains ated with liver failure is approximately 5% [1]. supplementary material, which is available to authorized users. Department of Pathology and Clinical Laboratory, National & Naoto Gotohda Cancer Center Hospital East, Kashiwa, Japan ngotohda@east.ncc.go.jp Course of Advanced Clinical Research of Cancer, Juntendo 1 University Graduate School of Medicine, Tokyo, Japan Department of Hepatobiliary and Pancreatic Surgery, National Cancer Center Hospital East, 6-5-1 Kashiwa-no-ha, Kashiwa, Chiba 277-8577, Japan Department of Diagnostic Radiology, National Cancer Center Hospital East, Kashiwa, Japan 123 Hepatology International (2018) 12:368–376 369 Preoperative precise assessment of liver function, extent of were used to analyze the correlations between the LSR and resection, and estimated remnant liver volume is important grading systems of liver function or laboratory data (patient to minimize the risks associated with liver surgery. cohort 2). Finally, a subgroup of all 112 patients who Gadolinium ethoxybenzyl diethylenetriamine pen- underwent liver resection was used to analyze the corre- taacetic acid (Gd-EOB-DTPA) is a liver-specific magnetic lations between the LSR and histological parameters (pa- resonance imaging (MRI) contrast agent. MRI using Gd- tient cohort 3). The patient flow chart of this study is shown EOB-DTPA (EOB-MRI) is widely used to detect and in Fig. 1. characterize hepatocellular carcinoma or metastatic liver tumors [2]. Gd-EOB-DTPA has a well-known metabolic MRI protocol pathway, and several studies have suggested that EOB- MRI is useful to evaluate liver function [3, 4]. The liver-to- All MRI studies were performed using one of the two 3.0 T spleen signal intensity ratio (LSR) on EOB-MRI has been scanners at our institution (Achieva or Ingenia, Philips used as a parameter to assess liver function [5–8], but the Medical Systems, Amsterdam, Netherlands). Contrast-en- conventional method to measure the LSR using two-di- hanced 3D fat-suppressed T1-weighted images were mensional (2D) regions of interest might be affected by obtained 20 min after intravenous administration of Gd- sampling errors or inter-evaluator variability [9]. EOB-DTPA for hepatobiliary phase imaging, using the In a related field, three-dimensional (3D) volumetric following parameters: repetition time (TR) 4 ms, echo time analysis systems are being used to simulate liver anatomy (TE) 2 ms, flip angle 10, slice thickness 4.6 mm, matrix for surgery. Using these systems, inter-evaluator variability size 512 9 512 (Achieva), and TR 3 ms, TE 2 ms, flip is expected to be reduced, as signal intensity from the angle 10, slice thickness 4.6 mm, matrix size 480 9 480 whole liver can be included in a semi-automatic analysis. (Ingenia). Gd-EOB-DTPA was administered at a dose of There have been no previous clinical studies in which the about 0.1 mL/kg, by rapid intravenous bolus injection 3D volumetric analysis system was used to measure the using a power injector (SONIC SHOT GX, NEMOTO- LSR on EOB-MRI. Therefore, the objective of this study KYORINDO, Tokyo, Japan), at a rate of 2 mL/s. was to evaluate the variability of the calculated LSR of EOB-MRI using a 3D volumetric analysis system, and to MRI data analysis investigate the correlations between the LSR and liver function parameters or histological findings. The LSR was calculated using images from the 20 min- delayed hepatobiliary phase, with the 3D volumetric analysis system SYNAPSE VINCENT (Fujifilm Medical, Materials and methods Tokyo, Japan), by a single investigator (Ma.Ku.), under supervision of an experienced radiologist (T.K.). First, the Patients investigator placed small operator-defined volumes of interest (VOIs), one in the liver and another in the spleen Between August 2014 and December 2015, a total of 304 parenchyma, avoiding vessels and tumors (Fig. 2a). Sec- consecutive patients underwent laboratory and radiological ond, the liver and spleen parenchyma were semi-automat- examinations as preoperative evaluations in consideration ically extracted using the image-processing algorithm of liver surgery at the National Cancer Center Hospital (Fig. 2b). Finally, the LSR was calculated as the average East, Japan. Of the 304 patients, 123 were excluded for the liver parenchyma signal intensity (Fig. 2c) divided by the following reasons: contraindications to EOB-MRI average spleen parenchyma signal intensity. Using the (n = 29), inconsistent MRI acquisition technique (n = 52), patient cohort 1, the variability of the calculated LSR incomplete laboratory data (n = 3), or prior splenectomy among four evaluators (surgeons with 7–11 years of clin- (n = 1). Thirty-eight patients who had undergone portal ical experience) was analyzed. Then, since inclusion of vein embolization were also excluded because the degree extrahepatic parenchymal tissue, such as portal and hepatic of liver enhancement from Gd-EOB-DTPA could depend veins, intrahepatic bile ducts, cysts, and tumors, might considerably on the portal vein flow. affect the average liver signal intensity, these structures The remaining 181 patients who underwent EOB-MRI were subtracted manually by the investigator (Fig. 2e). of the liver using a standardized imaging technique were Additionally, the vascular subtraction LSR (vsLSR) was used for the analysis in this study. Among them, 24 con- calculated, and its correlation with the LSR was analyzed secutive patients who underwent preoperative evaluation using cohort 1 (Fig. 1). from November 2015 to December 2015 were used to examine the variability of the LSR among four different evaluators (patient cohort 1). Then, all the 181 patients 123 370 Hepatology International (2018) 12:368–376 Fig. 1 Flow-diagram describing the patient cohorts in this study. Gd-EOB-DTPA gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid; LSR liver-to-spleen signal intensity ratio Grading systems of liver function and laboratory stained by hematoxylin and eosin (HE), azo carmine ani- line blue (AZAN), and smooth muscle actin (SMA) data (Fig. 3). The specimen slides were scanned using the The following laboratory parameters were obtained preop- NanoZoomer 2.0 system (Hamamatsu Photonics, Hama- matsu, Japan), and morphometric analysis was performed eratively and within a month before or after EOB-MRI were collected from the patients’ medical records: indocyanine using the WinROOF 6.5 image processing software pro- gram (MITANI Corporation, Tokyo, Japan). Histological green retention rate at 15 min (ICGR15), white blood cell count, platelet count, prothrombin activity, and serum levels analyses were performed by a single investigator (Ma.Ku.), under supervision of an experienced pathologist (Mo.Ko.). of hemoglobin, albumin, total bilirubin, aspartate amino- transferase, alanine aminotransferase, and creatinine. Grad- Liver fibrosis was assessed in three ways. First, using HE-stained slides, the fibrosis stage was morphologically ing systems of liver function, such as the Child–Pugh score and the albumin-bilirubin (ALBI) score, were analyzed for categorized by the METAVIR scoring system as follows: F0, no fibrosis; F1, portal fibrosis without septa; F2, few the correlation with the LSR. The Child–Pugh score was graded into A, B, or C, using five variables, including septa; F3, numerous septa without cirrhosis; F4, cirrhosis bilirubin, albumin, prothrombin, ascites status, and degree of [10]. Second, the ratio of fibrosis (ROF) on AZAN-stained slides (ROFazan) was quantified using morphometric encephalopathy. The ALBI score was calculated as follows: analysis from a color-detecting algorithm (Fig. 3b), using 0:085 ðalbumin g/lÞþ 0:66 the same technique as reported in a previous study [11]. logðtotal bilirubin lmol/lÞ; Third, the ROF on SMA-stained slides (ROFsma) was and categorized into the following three grades: quantified in a similar manner (Fig. 3d). Liver steatosis was also assessed in two ways. First, the \  2:60; ALBI 1; [ 2:06 to  1:39; ALBI 2; steatosis grade was morphologically categorized as in [  1:39; ALBI 3: Kleiner et al. [12]: grade 0, \ 5%; grade 1, 5–33%; grade 2, 34–66%; grade 3, [ 67%. Second, the ratio of droplet Using patient cohort 2, the correlations between the LSR steatosis (ROS) was quantified using morphometric anal- and grading systems of liver function or other laboratory ysis from the color-detecting algorithm (Fig. 3f). data were analyzed (Fig. 1). Statistical analysis Histological analysis of the surgical specimen Statistical analyses were performed using JMP (version From the 181 studied patients, 112 patients underwent liver 12.0.10; SAS Institute, Cary, NC). Ebel’s intraclass cor- resection (patient cohort 3) (Fig. 1). Their surgical speci- relation coefficients were used to evaluate the inter-eval- mens of non-tumoral liver tissues were fixed in formalin uator variability of the LSR and the correlation between the and embedded in paraffin. Then, 4-lm-thick sections were 123 Hepatology International (2018) 12:368–376 371 Fig. 2 Image analysis using the three-dimensional volumetric system. calculated semi-automatically. d Liver parenchyma extracted before a The investigator placed a small volume of interest (VOI) in the liver subtracting vascular structures. e Liver parenchyma after subtracting parenchyma. b The liver parenchyma was semi-automatically vascular structures, such as portal veins and hepatic veins, and extracted from the initial VOI seed. c The three-dimensional liver intrahepatic bile ducts volume was extracted, and the average liver signal intensity was LSR and vsLSR in patient cohort 1. Correlations between hepatocellular carcinoma in 94 patients (50%), metastatic the LSR and the clinicopathological factors were assessed liver cancer in 76 patients (42%), perihilar cholangiocar- by the standard Pearson’s correlation coefficient or cinoma in five patients (3%), cholangiocellular carcinoma Spearman’s rank correlation coefficient in patient cohorts 2 in four patients (2%), and other disease in six patients and 3. The correlations between the LSR and fibrosis stage (3%). The primary lesion of metastatic liver cancer was or steatosis grade were evaluated using pairwise compar- colon in 71 patients, stomach in two patients, pancreas in isons with the Mann–Whitney test in patient cohort 3. Two- one patient, lung in one patient, and biliary tract in one sided p values of less than 0.05 were considered indicative patient, respectively. There was an underlying liver infec- of significance. tion of the hepatitis C virus in 44 patients (24%) and of the hepatitis B virus in nine patients (5%). Of the 181 patients undergoing EOB-MRI of the liver in this study, 177 Results (97.8%) patients were classified as having the Child–Pugh grade A and four (2.2%) patients as having the Child–Pugh Patients’ demographics grade B. The median value of ICGR15 was 13.1% (range, 2.9–58.9%). The therapies provided for the liver tumors The median age was 70 (range, 39–90), and 130 (72%) of were as follows: surgical resection in 112 patients (62%), 181 patients were male. The diagnoses were as follows: chemotherapy in 26 patients (14%), radiofrequency 123 372 Hepatology International (2018) 12:368–376 Fig. 3 Areas of fibrosis and steatosis were calculated using morphometric analysis of color- detecting algorithm (WinROOF software, version 6.5; MITANI Corporation, Tokyo, Japan). a Azo carmine aniline blue (AZAN) and c smooth muscle actin (SMA) stain. The AZAN/ SMA-positive area was determined using a color- detecting algorithm and is represented as bright green in b and d. The ratio of fibrosis (ROF) was calculated as the percentage area of the entire field and the AZAN/SMA- positive area. Hematoxylin– eosin stain of fatty liver (e). Fat droplets were determined using a color-detecting algorithm and are represented as bright green in image (f). The ratio of steatosis (ROS) was calculated as the percentage area of the entire field and the fat droplets area ablation in 11 patients (6%), transcatheter arterial Correlation between the LSR and grading chemoembolization in 13 patients (7%), proton beam systems of liver function (patient cohort 2, radiation therapy in 12 patients (7%), and supportive care n = 181) in seven patients (4%). Correlations between the LSR and grading systems of liver Inter-evaluator variability of the LSR function were summarized in Table 1. The mean LSR was and correlation between the LSR and vsLSR lower in the patients with the Child–Pugh grade B than (patient cohort 1, n = 24) those with the Child–Pugh grade A (1.57 vs 1.97, p = 0.014). The mean LSR was lower in the patients with The intraclass correlation coefficient of the LSR amongst the ALBI score 2 than those with the ALBI score 1 (1.64 vs the four evaluators calculated using the 3D volumetric 2.05, p \ 0.001). analysis system, was 0.986. The intraclass correlation coefficient between the LSR and vsLSR, evaluated by a single investigator, was 0.987. 123 Hepatology International (2018) 12:368–376 373 Table 1 Correlations between the liver-to-spleen ratio (LSR) and aminotransferase (r = - 0.422, p \ 0.001), and alanine grading systems of liver function (patient cohort 2, n = 181) aminotransferase (r = - 0.287, p \ 0.001). n = 181 LSR (mean) p Correlations between the LSR and histological Child–Pugh score findings (patient cohort 3, n = 112) A 177 (97.8%) 1.97 = 0.014 B 4 (2.2%) 1.57 Correlations between the LSR and histological findings, ALBI score including liver fibrosis and steatosis, are summarized in 1 142 (78.5%) 2.05 \ 0.001 Table 2 and Fig. 4. In terms of liver fibrosis, negative 2 39 (21.5%) 1.64 correlations were observed between the LSR and the METAVIR score (r = - 0.556, p \ 0.001), between the LSR and ROFazan (r = - 0.424, p \ 0.001), and between the LSR and ROFsma (r = - 0.592, p \ 0.001) (Table 2). Correlations between the LSR and laboratory The LSR value was significantly greater for fibrosis stages data (patient cohort 2, n = 181) F0 and F1 than for stages F2, F3, or F4 (each pairwise comparison, p \ 0.001) (Fig. 4a). In terms of liver Correlations between the LSR and laboratory data are steatosis, negative correlations were observed between the summarized in Table 2. Positive correlations were LSR and the Kleiner’s grade (r = - 0.396, p \ 0.001) and observed between the LSR and the following parameters: between the LSR and ROS (r = - 0.428, p \ 0.001). The platelet count (r = 0.307, p \ 0.001), serum level of LSR value was significantly greater for steatosis grade 0 albumin (r = 0.453, p \ 0.001), and prothrombin activity than for grade 1 (p \ 0.001), grade 2 (p \ 0.001), or grade (r = 0.426, p \ 0.001). Negative correlations were 3 (each pairwise comparison, p = 0.001) (Fig. 4b). observed between the LSR and the following parameters: ICGR15 (r = - 0.601, p \ 0.001) and serum levels of total bilirubin (r = - 0.370, p \ 0.001), aspartate Discussion Table 2 Correlations between the liver-to-spleen ratio (LSR) and In this study, the clinical utility of the LSR, calculated as laboratory data or histological findings the signal intensity of the liver parenchyma divided by that rp of the spleen on EOB-MRI, was evaluated using a 3D volumetric analysis system. The reproducibility of the Laboratory data (patient cohort 2, n = 181) calculated LSR amongst the different evaluators was very ICGR15 - 0.601 \ 0.001 high. The LSR was shown to be correlated with grading White blood cell 0.067 0.368 systems of liver function, such as the Child–Pugh score and Platelet count 0.307 \ 0.001 the ALBI score. In addition, significant correlations Prothrombin activity 0.426 \ 0.001 between the LSR and ICGR15 and between the LSR and Hemoglobin 0.021 0.783 histological findings of liver fibrosis or steatosis were Albumin 0.453 \ 0.001 observed. This is the first report using a 3D volumetric Total bilirubin - 0.370 \ 0.001 analysis system of EOB-MRI to evaluate liver function. Aspartate aminotransferase - 0.422 \ 0.001 Conventionally, signal intensity of the liver parenchyma Alanine aminotransferase - 0.287 \ 0.001 on EOB-MRI has been measured by ROIs arbitrarily Creatinine - 0.116 0.120 selected on a few 2D MRI slices. However, such systems Histological findings (patient cohort 3, n = 112) might include a certain degree of selection bias of the Liver fibrosis ROIs. On the other hand, 3D volumetric analysis systems, METAVIR score - 0.556 \ 0.001 which extract the entire volume of specific internal organs ROF (AZAN stain) - 0.424 \ 0.001 based on high-precision image processing algorithm, have ROF (SMA stain) - 0.592 \ 0.001 been introduced in clinical settings in recent years. The Liver steatosis benefits of preoperative simulation of liver anatomy with Kleiner grade - 0.396 \ 0.001 the use of 3D image visualization technologies have been ROS - 0.428 \ 0.001 shown in several papers [13, 14]. Takamoto et al. showed a high correlation between estimated liver volume on pre- ICGR15 indocyanine green retention rate at 15 min; ROF ratio of fibrosis; ROS ratio of steatosis; AZAN azo carmine aniline blue; SMA operative computed tomography and the weight of the smooth muscle actin resected specimens [15]. Ogawa et al. reported the utility of a 3D volumetric analysis system on computed tomography 123 374 Hepatology International (2018) 12:368–376 Fig. 4 Box plots of the liver-to-spleen ratio (LSR) according to the fibrosis stage (a) and steatosis grade (b). *Significant differences in pairwise comparisons using the Mann–Whitney test images to predict the area of the liver that is embolized the strongest correlation between the LSR and ICGR15, after transcatheter arterial chemoembolization [16]. The compared to the other laboratory data analyzed. advantage of 3D volumetric analysis systems compared to There are several studies showing a significant correla- 2D ones is that they enable the calculation of the average tion between signal intensity of the liver on 2D analysis whole liver signal intensity automatically, from only a systems on EOB-MRI and liver fibrosis [6, 20–22]. These small VOI drawn in the liver parenchyma. The present studies evaluated liver fibrosis using either the METAVIR study showed that the inter-evaluator correlation coeffi- score or the New INUYAMA classification as categorical cient of the LSR values was 0.986. This means that the variables assessed by pathologists. In this study, morpho- LSR, calculated using a 3D volumetric analysis system of metric analysis, using image processing software, enabled EOB-MRI, is an objective, precise, and quantitative index the quantification of fibrosis or steatosis in the liver, and to measure liver signal intensity. showed a significantly inverse correlation between the LSR There are several studies that used 2D analysis systems and fibrosis or steatosis. These data suggest that infiltration on EOB-MRI to evaluate liver function. These studies in the liver parenchyma of stromal tissues, such as fibrosis indicated that liver signal intensity, measured from ROIs, or steatosis, decreases the relative area of hepatocytes and was correlated with liver functional reserve markers, such leads to the reduction of signal intensity of the whole liver as ICGR15, prothrombin activity, or liver fibrosis on EOB-MRI. Interestingly, as shown in the Supplemen- [3–5, 17]. The present study showed that the LSR calcu- tary Table, ICGR15 was not significantly correlated with lated using a 3D volumetric analysis system on EOB-MRI fibrosis or steatosis in the liver. These results suggest that was correlated with grading systems of liver function and the function of the liver that was affected by fibrosis or laboratory data, such as platelet count, serum levels of steatosis was not reflected accurately by ICGR15 but by the albumin, total bilirubin, aspartate aminotransferase, alanine signal intensity of the whole liver on EOB-MRI. Therefore, aminotransferase, prothrombin activity, and ICGR15. the LSR, calculated by a 3D volumetric analysis system on ICGR15 is one of the well-known biochemical indices used EOB-MRI, may indicate liver function more accurately to predict post-hepatectomy remnant liver function. In than the other liver function parameters. There might be principle, ICG is taken up to hepatocytes via the organic concerns regarding other components in the liver than anion transporting polypeptide or Na -taurocholate co- normal parenchyma, fibrosis, and steatosis. Majority of transporting polypeptides located in the sinusoidal mem- previous studies reported no significant correlations brane, and is excreted into the biliary system via the ATP- between fibrosis or steatosis and iron deposition dependent export pump multidrug resistance-associated histopathologically or radiographically [23–25]. Moreover, protein 2 without biotransformation [18]. Gd-EOB-DTPA there were no patients with hemosiderosis or hemochro- is also taken up to hepatocytes via the organic anion matosis clinically in our patient cohort. Therefore, liver transporting polypeptides, and is excreted into the biliary fibrosis or steatosis was considered to be evaluated using system via the multidrug resistance-associated protein 2 EOB-MRI regardless of the iron deposition. [19]. Therefore, in the light of the metabolic mechanism In this study, there were five patients with perihilar theory, the signal intensity of the liver on EOB-MRI would cholangiocarcinoma who developed obstructive jaundice at be expected to reflect ICGR15. The present study showed presentation. However, we consider that the influence of 123 Hepatology International (2018) 12:368–376 375 obstructive jaundice on evaluating liver function using In conclusion, the LSR calculated using a 3D volumetric EOB-MRI was expected to be minimal, because all the analysis system on EOB-MRI were highly reproducible, patients who had obstructive jaundice were re-examined by and were correlated with grading systems of liver function, EOB-MRI for operative indication after biliary drainage laboratory data, and histological findings. EOB-MRI using and resolution of jaundice. However, evaluation of liver a 3D volumetric analysis system may be a reliable modality function using EOB-MRI for the patients with obstructive to evaluate liver function. jaundice may be a subject of future study. Acknowledgements The authors would like to thank Ryo Morisue, Our study has several limitations. First, there was a Toshiyuki Suzuki and Hidetoshi Aizawa for assisting with data col- considerable number of patients (52/304, 17%) that had to lection, and Kenichi Ikejima M.D. (Department of Gastroenterology, be excluded due to a different EOB-MRI acquisition pro- Juntendo University Graduate School of Medicine, Tokyo, Japan) for tocol (mostly due to a different slice thickness). During the his technical support. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit study period, the slice thickness of 4.6 mm might be rel- sectors. atively thick for 3D volumetric analysis. The MRI protocol was determined to diagnose liver tumors but not to evaluate Compliance with ethical standards liver function in this study. However, our study showed that the LSR was correlated with the Child–Pugh grade, the Conflict of interest Masashi Kudo, Naoto Gotohda, Motokazu Sugi- ALBI score, ICGR15 and other laboratory data, and his- moto, Tatsushi Kobayashi, Motohiro Kojima, Shinichiro Takahashi, Masaru Konishi, and Ryuichi Hayashi declare that they have no tological findings, although the MRI conditions were set to conflicts of interest. diagnose liver tumors. These results suggest that EOB-MRI would be applied for evaluating liver function in the clin- Ethical approval This study protocol conformed to the ethical ical situation. Furthermore, the MRI protocol in the pre- guidelines of the 1975 Declaration of Helsinki and was approved by vious studies that evaluated liver function using EOB-MRI the institutional review board of the National Cancer Center, Japan. did not differ widely from that in our study: the slice Informed consent Informed consent was obtained from each patient thickness was 5.0 mm in the study by Okada et al. [8], included in the study. 4.0 mm by Nishie et al. [7], and 3.8 mm by Matsushima et al. [4]. Second, in the 3D volumetric analysis, extra- Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creative hepatic-parenchymal tissues such as portal vein or hepatic commons.org/licenses/by/4.0/), which permits unrestricted use, dis- vein were included in the whole liver signal intensity cal- tribution, and reproduction in any medium, provided you give culation: the LSR was the metric primarily used. However, appropriate credit to the original author(s) and the source, provide a since the study showed a high correlation between the LSR link to the Creative Commons license, and indicate if changes were made. and vsLSR (LSR excluding extrahepatic-parenchymal tis- sues), and demonstrated that the LSR adequately reflects contrast enhancement of the liver parenchyma, this sug- gests that liver function can be evaluated without sub- References tracting vessels and vascular perfusion areas. Third, two different MR scanners were used in this study. This might 1. Yadav K, Shrikhande S, Goel M. Post hepatectomy liver failure: have caused a systemic bias in the calculation of liver or concept of management. J Gastrointest Cancer. 2014;45(4):405–413. https://doi.org/10.1007/s12029-014-9646-3 spleen signal intensities. However, the difference is not Epub 2014/08/12. assumed to be significant, as the MRI protocols were 2. Huppertz A, Balzer T, Blakeborough A, Breuer J, Giovagnoni A, similar between the two different machines and all images Heinz-Peer G, et al. Improved detection of focal liver lesions at were taken by T-1 weighted technique. In a previous MR imaging: multicenter comparison of gadoxetic acid-enhanced MR images with intraoperative findings. Radiology. multicenter study, Okada et al. reported no significant 2004;230(1):266–275. https://doi.org/10.1148/radiol.2301020269 differences in signal intensity of the liver on EOB-MRI Epub 2003/12/26. between different MRI scanners [8]. Fourth, there were a 3. Utsunomiya T, Shimada M, Hanaoka J, Kanamoto M, Ikemoto T, large number of patients excluded for having prior history Morine Y, et al. Possible utility of MRI using Gd-EOB-DTPA for estimating liver functional reserve. J Gastroenterol. of portal vein embolization. Previous studies showed a 2012;47(4):470–476. https://doi.org/10.1007/s00535-011-0513-8 significant decrease of signal intensity of the liver lobe Epub 2011/12/16. after portal vein embolization [26]. Therefore, in the pre- 4. Matsushima S, Sato Y, Yamaura H, Kato M, Kinosada Y, Era S, sent study, we decided to investigate the entire liver et al. Visualization of liver uptake function using the uptake contrast-enhanced ratio in hepatobiliary phase imaging. Magn function, instead of regional liver function. Preoperative Reson Imaging. 2014;32(6):654–659. https://doi.org/10.1016/j. evaluation of the regional liver function using EOB-MRI mri.2014.02.017 Epub 2014/03/19. for major hepatectomy would be the subject of future study. 123 376 Hepatology International (2018) 12:368–376 5. Motosugi U, Ichikawa T, Sou H, Sano K, Tominaga L, Kitamura 16. Ogawa C, Minami Y, Morita M, Noda T, Arasawa S, Izuta M, T, et al. Liver parenchymal enhancement of hepatocyte-phase et al. Prediction of embolization area after conventional tran- images in Gd-EOB-DTPA-enhanced MR imaging: which bio- scatheter arterial chemoembolization for hepatocellular carci- logical markers of the liver function affect the enhancement? noma using SYNAPSE VINCENT. Dig Dis. J Magn Reson Imaging. 2009;30(5):1042–1046. https://doi.org/ 2016;34(6):696–701. https://doi.org/10.1159/000448859 Epub 10.1002/jmri.21956 Epub 2009/10/27. 2016/10/18. 6. Motosugi U, Ichikawa T, Oguri M, Sano K, Sou H, Muhi A, et al. 17. Yamada A, Hara T, Li F, Fujinaga Y, Ueda K, Kadoya M, et al. Staging liver fibrosis by using liver-enhancement ratio of Quantitative evaluation of liver function with use of gadoxetate gadoxetic acid-enhanced MR imaging: comparison with aspartate disodium-enhanced MR imaging. Radiology. aminotransferase-to-platelet ratio index. Magn Reson Imaging. 2011;260(3):727–733. https://doi.org/10.1148/radiol.11100586 2011;29(8):1047–1052. https://doi.org/10.1016/j.mri.2011.05.007 Epub 2011/06/30. Epub 2011/07/22. 18. Hoekstra LT, de Graaf W, Nibourg GA, Heger M, Bennink RJ, 7. Nishie A, Ushijima Y, Tajima T, Asayama Y, Ishigami K, Stieger B, et al. Physiological and biochemical basis of clinical Kakihara D, et al. Quantitative analysis of liver function using liver function tests: a review. Ann Surg. 2013;257(1):27–36. superparamagnetic iron oxide- and Gd-EOB-DTPA-enhanced https://doi.org/10.1097/sla.0b013e31825d5d47 Epub 2012/07/28. MRI: comparison with technetium-99 m galactosyl serum albu- 19. Van Beers BE, Pastor CM, Hussain HK. Primovist, eovist: what min scintigraphy. Eur J Radiol. 2012;81(6):1100–1104. https:// to expect? J Hepatol. 2012;57(2):421–429. https://doi.org/10. doi.org/10.1016/j.ejrad.2011.02.053 Epub 2011/03/26. 1016/j.jhep.2012.01.031 Epub 2012/04/17. 8. Okada M, Murakami T, Kuwatsuru R, Nakamura Y, Isoda H, 20. Watanabe H, Kanematsu M, Goshima S, Kondo H, Onozuka M, Goshima S, et al. Biochemical and clinical predictive approach Moriyama N, et al. Staging hepatic fibrosis: comparison of and time point analysis of hepatobiliary phase liver enhancement gadoxetate disodium-enhanced and diffusion-weighted MR on Gd-EOB-DTPA-enhanced MR images: a multicenter study. imaging—preliminary observations. Radiology. Radiology. 2016;281(2):474–483. https://doi.org/10.1148/radiol. 2011;259(1):142–150. https://doi.org/10.1148/radiol.10100621 2016151061 Epub 2016/05/20. Epub 2011/01/21. 9. Imura S, Shimada M, Utsunomiya T. Recent advances in esti- 21. Nojiri S, Kusakabe A, Fujiwara K, Shinkai N, Matsuura K, Iio E, mating hepatic functional reserve in patients with chronic liver et al. Noninvasive evaluation of hepatic fibrosis in hepatitis C damage. Hepatol Res. 2015;45(1):10–19. https://doi.org/10.1111/ virus-infected patients using ethoxybenzyl-magnetic resonance hepr.12325 Epub 2014/03/13. imaging. J Gastroenterol Hepatol. 2013;28(6):1032–1039. https:// 10. Bedossa P, Poynard T. An algorithm for the grading of activity in doi.org/10.1111/jgh.12181 Epub 2013/02/26. chronic hepatitis C. The METAVIR cooperative study group. 22. Shimamoto D, Nishie A, Asayama Y, Ushijima Y, Takayama Y, Hepatology. 1996;24(2):289–293. https://doi.org/10.1002/hep. Fujita N, et al. MR prediction of liver function and pathology 510240201 Epub 1996/08/01. using Gd-EOB-DTPA: effect of liver volume consideration. 11. Sugimoto M, Takahashi S, Kojima M, Kobayashi T, Gotohda N, Biomed Res Int. 2015;2015:141853. https://doi.org/10.1155/ Konishi M. In patients with a soft pancreas, a thick parenchyma, a 2015/141853 Epub 2015/11/27. small duct, and fatty infiltration are significant risks for pancreatic 23. Franca M, Alberich-Bayarri A, Marti-Bonmati L, Oliveira P, fistula after pancreaticoduodenectomy. J Gastrointest Surg. 2017. Costa FE, Porto G, et al. Accurate simultaneous quantification of https://doi.org/10.1007/s11605-017-3356-7. liver steatosis and iron overload in diffuse liver diseases with 12. Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, MRI. Abdom radiol (New York). 2017;42(5):1434–1443. https:// Cummings OW, et al. Design and validation of a histological doi.org/10.1007/s00261-017-1048-0 Epub 2017/01/23. scoring system for nonalcoholic fatty liver disease. Hepatology. 24. Szurowska E, Sikorska K, Izycka-Swieszewska E, Nowicki T, 2005;41(6):1313–1321. https://doi.org/10.1002/hep.20701 Epub Romanowski T, Bielawski KP, et al. The role of MR imaging in 2005/05/26. detection of hepatic iron overload in patients with cirrhosis of 13. Ohshima S. Volume analyzer SYNAPSE VINCENT for liver different origins. BMC Gastroenterol. 2010;10:13. https://doi.org/ analysis. J Hepatobiliary Pancreat Sci. 2014;21(4):235–238. 10.1186/1471-230x-10-13 Epub 2010/01/29. https://doi.org/10.1002/jhbp.81 Epub 2014/02/13. 25. Ludwig J, Batts KP, Moyer TP, Baldus WP, Fairbanks VF. Liver 14. Mochizuki K, Takatsuki M, Soyama A, Hidaka M, Obatake M, biopsy diagnosis of homozygous hemochromatosis: a diagnostic Eguchi S. The usefulness of a high-speed 3D-image analysis algorithm. Mayo Clin Proc. 1993;68(3):263–267 Epub 1993/03/ system in pediatric living donor liver transplantation. Ann 01. Transplant. 2012;17(1):31–34 Epub 2012/04/03. 26. Geisel D, Raabe P, Ludemann L, Malinowski M, Stockmann M, 15. Takamoto T, Hashimoto T, Ogata S, Inoue K, Maruyama Y, Seehofer D, et al. Gd-EOB-DTPA-enhanced MRI for monitoring Miyazaki A, et al. Planning of anatomical liver segmentectomy future liver remnant function after portal vein embolization and and subsegmentectomy with 3-dimensional simulation software. extended hemihepatectomy: a prospective trial. Eur Radiol. Am J Surg. 2013;206(4):530–538. https://doi.org/10.1016/j.amj 2016;27(7):3080–3087. https://doi.org/10.1007/s00330-016- surg.2013.01.041 Epub 2013/07/03. 4674-y Epub 2016/12/15.

Journal

Hepatology InternationalSpringer Journals

Published: Jun 2, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off