Rules, Roles, and Room for Discussion in Gadoxetic Acid-Enhanced Magnetic Resonance Liver Imaging: Current Knowledge and Future Challenges

Magn Reson Med Sci, Vol. 12, No. 3, pp. 161–175, 2013 doi:10.2463/mrms.2012-0085 2013 Japanese Society for Magnetic Resonance in Medicine

REVIEW

Rules, Roles, and Room for Discussion in Gadoxetic Acid-enhanced Magnetic Resonance Liver Imaging: Current Knowledge and Future Challenges

Utaroh MOTOSUGI*, Tomoaki ICHIKAWA, and Tsutomu ARAKI

Department of Radiology, University of Yamanashi 1110 Shimokato, Chuo-shi, Yamanashi 409–3898, Japan (Received December 3, 2012; Accepted April 3, 2013; published online July 12, 2013)

Since its launch onto the world medical market, gadoxetic acid has received a great deal of attention from a diverse group of researchers, including gastroenterologists, hepato- logists, and abdominal radiologists. Gadoxetic acid in liver magnetic resonance (MR) imaging might be considered to have excellent utility. However, several issues remain that radiologists should keep in mind when utilizing this product. We compiled this review to introduce gadoxetic acid, describe how it should be used, highlight the situations in which it is most and least reliable, and suggest potential avenues for future research.

Keywords: gadoxetic acid, liver, magnetic resonance imaging

now available for MR imaging of the liver-extracel- Introduction lular gadolinium-based contrast agents (GBCAs) Gadoxetic acid, or gadoxetate disodium (Gd- and liver-speciˆc contrast agents, such as gadoxetic EOB-DTPA), is a recently launched liver-speciˆc acid (Primovist or Eovist, Bayer Healthcare, Ger- magnetic resonance (MR) contrast agent, distin- many) and gadobenate dimeglumine (Gd-BOPTA). guished by its capacity for rapid (º30 min) evalua- Frydrychowicz and associates compared these lat- tion of both dynamic contrast-enhanced phase im- ter 2 products extensively in a recent paper,1 and ages and liver-speciˆc phase images. here, we focus exclusively on the characteristics The use of gadoxetic acid in liver imaging has that should be considered when choosing between been the focus of a variety of studies, and in the gadoxetic acid and conventional extracellular GB- current review, we aim to summarize the ˆndings of CAs when performing MR imaging of the liver. that work to describe how gadoxetic acid can and We believe that gadoxetic acid should be the ˆrst should be used. We describe how to use gadoxetic choice for liver imaging in most cases, but we acid during clinical MR examinations and focus on should also be aware of the major disadvantages of the results of many parameter optimization studies gadoxetic acid. First, its use precludes liver equilib- that aimed to standardize gadoxetic acid adminis- rium phase because signal intensity is always higher tration and scanning protocols during MR imaging for the liver parenchyma than for vessels after the of the liver, review the literature on the diagnostic portal-venous phase. This makes it di‹cult to diag- ability of gadoxetic acid, and outline potential fu- nose hemangioma, because pooling of contrast a- ture uses of gadoxetic acid and issues that require gent, a key feature of hemangioma, cannot be ob- 2 further clinical and experimental evaluation. served. Very high signal intensity on a T2-weighted image might help distinguish between hemangioma and liver malignancies, such as metastasis,3 but Rules Guiding Use of Gadoxetic Acid gadoxetic acid should not be used when the purpose Gadoxetic acid or extracellular gadolinium-based of the MR examination is to conˆrm a hemangioma contrast agents? diagnosis rather than detect small focal liver le- In most countries, 2 types of contrast agent are sions. The second disadvantage of gadoxetic acid is that its uptake is not always su‹cient to provide the *Corresponding author, Phone: ＋81-55-273-1111, Fax: ＋81- contrast necessary to distinguish between liver pa- 55-273-6744, E-mail: Utaroh-motosugi＠nifty.com renchyma and lesions4–6 (Fig. 1a, b). If the liver

161 162 U. Motosugi et al.

Fig. 1. Gadoxetic acid-enhanced hepatocyte phase magnetic resonance (MR) images in patients with cirrhosis with su‹cient (a) and insu‹cient (b) liver enhancement. Precise detection of small liver lesions requires a contrast enhance-

ment ratio, or ratio of signal intensity between the liver and spleen (SILiver/ SISpleen), greater than 1.5. Hepatocellular carcinomas (HCCs) are well recognized in the liver (arrows), whereas small HCC is ambiguous in the liver with insu‹cient liver enhancement (arrowhead). is not well enhanced on hepatocyte phase images, vide visually equivalent enhancements. thereisnoreasontousegadoxeticacidforliverim- aging. Several clinical parameters, including Child- Phase names Pugh score, serum bilirubin levels, and indocyanine When using extracellular GBCAs, dynamic green test results, are associated with uptake of phases include the arterial or hepatic arterial phase gadoxetic acid (or signal intensity of the liver on followed by the portal venous phase, which is ob- hepatocyte phase images).4–7 No known imaging tained about 60 to 70 s after injection; enhance- parameter predicts liver enhancement on hepato- ment of the liver parenchyma is highest during the cyte phase images before gadoxetic acid adminis- portal venous phase. Three to 5 min after injection, tration,7 though liver ˆbrosis or stiŠness as meas- the equilibrium, delayed, or late phase begins, dur- ured by MR elastography might indicate insuf- ing which the signal intensity of the liver parenchy- ˆcient liver enhancement.8 We suggest using an ma decreases to the level found in the vessels. When alternative contrast agent, such as an extracellular gadoxetic acid is used as the contrast agent, the sig- GBCA, when previous MR examinations indicate nal intensity of the liver continuously increases un- insu‹cient liver enhancement after gadoxetic acid til it plateaus 20 min after the injection.12 During administration. the hepatocyte or hepatobiliary phase, 10 to 20 min Gadoxetic acid contains half the amount of after injection, contrast between the liver parenchy- gadolinium found in extracellular GBCAs and is ma and vessels or focal liver lesions is maximal. administered in doses that are half those of GB- Most researchers use the terms arterial phase and CAs; thus, examinations with gadoxetic acid utilize portal venous phase regardless of whether gadoxet- only a quarter the amount of gadolinium (0.1 ic acid or extracellular GBCAs are used as the con- mmol/kg for extracellular GBCAs versus 0.025 trast agent. The term used to describe the phase be- mmol/kg for gadoxetic acid). Contrast enhance- tween the dynamic and hepatocyte phases can be ment of the vessels is lower on the arterial phase for controversial. The term equilibrium phase should gadoxetic acid-enhanced images than for extracel- be avoided because the concentration of gadoxetic 9 lular GBCAs. The T1-shortening eŠect of gadoxet- acid is consistently higher in the liver parenchyma ic acid is much larger than that of extracellular GB- than vessels, that is, the concentration of gadoxetic CAs, but that can result in equivalent or even great- acidisneveratequilibrium.Theconsensusreport er contrast enhancements in the aorta.10 A recent from the Second International Forum for Liver study revealed that use of a slow injection rate (one MRI used the term dynamic late phase to refer to mL/s) can facilitate equivalent arterial enhance- the mid-period 2 to 3 min after administration of ments between gadoxetic acid and extracellular GB- gadoxetic acid.13 CAs.11 Currently, no published data compare enhancement of liver lesions on arterial phase images Rules for the arterial phase enhanced by gadoxetic acid and extracellular GB- Optimal arterial phase images are required for CAs. In our experience, however, the 2 agents pro- MR imaging of the liver because many focal liver

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 163163 lesions show their own characteristic vascularity. useful.13,17 However, there is little information on Nonetheless, acquisition of arterial phase images of the optimal time delay after the arrival of the con- su‹cient quality may be di‹cult because methods trast agent to the aorta. Researchers have used of contrast agent injection and the very short op- delay times ranging from 7 to 15 s.17,20–22 If bolus timal time window for the arterial phase impact im- timing cannot be manipulated as described above, age quality. timing may instead be adjusted via the test-injec- Gradient echo sequence with 3-dimensional ac- tionmethod,inwhichasmallamount(e.g.,0.5 quisition (3D-GRE), which has a high spatial reso- mL) of contrast agent is injected before dynamic lution and signal-to-noise ratio, is usually used to study to measure the length of time for the agent to obtain gadoxetic acid-enhanced images.13 Ringing reach the aorta. artifacts sometimes appear on arterial phase images obtained with 3D-GRE because of k-space in- Rules for the hepatocyte phase homogeneity. Although the artifacts might include Delay before hepatocyte phase image distortion due to patient movement or res- The hepatocyte phase, also known as the hepato- piratory motion, the truncation artifact is sup- biliary phase, is the period during which the hepato- posed to play a major role in ringing artifacts.14 cytes take up most of the contrast agent and there is Truncation artifacts occur because of insu‹cient maximal contrast between the liver parenchyma sampling at high spatial frequencies and appear at and vasculature, including the portal and hepatic interfaces with instantaneous transitions from high veins. Classically, the hepatocyte phase has been to low signal intensity, such as the interface be- deˆned as beginning approximately 20 min after tween arteries with a high concentration of contrast contrast injection and lasting at least 2 hours.12,33 agent and surrounding soft tissue.15 In addition, In one Phase II study of gadoxetic acid, the tumor- truncation artifacts may result when an abrupt in- to-liver contrast plateaued 20 min after contrast in- crease in signal during signal acquisition from the jection, remained steady an additional 25 min.23 in‰ow of contrast agent into the artery being im- However, it is common to see hepatocyte phase im- aged changes the amplitude of the arterial signal in ages obtained only 10 to 20 min after contrast in- the k-space.16 jection.24–26 Indeed, the optimal hepatocyte phase can be deˆned as the period of time during which How to avoid k-space inhomogeneity there is su‹cient tumor-to-liver contrast or tumor Although a saline ‰ush is mandatory after bolus conspicuity. In other words, if tumor conspicuity is injection, use of a low injection rate will help appropriate, this phase may occur as early as 10 prolong the injection duration and decrease k-space min after injection,27 but 20 min after injection inhomogeneity.17,18 On the other hand, a high injec- may be too early to obtain images if optimal con- tion rate, such as 3 mL/s, results in short injection trast has not been achieved. In fact, in patients with duration—one purported cause of k-space inhomo- normal hepatic function, su‹cient liver enhance- geneity and truncation artifacts. Recent publica- ment can almost always be achieved 10 min after tions have recommended an injection rate of one injection;thus,itisnolongernecessarytowait20 mL/s to achieve su‹cient enhancement and im- min to perform imaging, even for small lesions.27–29 prove image quality.17,19 Use of a square matrix, slower injection rate, shorter scanning time, and se- Improving hepatocyte phase images quential view ordering are also reported to reduce Hepatocyte phase images are usually obtained artifacts.14 with a low ‰ip angle, such as 129or 159,asconven- tionally used in 3D-GRE sequences. Yet, a larger Acquiring images during the optimal arterial phase ‰ip angle may be optimal to maximize tumor-to- time window liver contrast because the presence of contrast The injection volume of gadoxetic acid, 0.1 agent can shorten T1 relaxation time in the liver mL/kg of body weight, is half that of extracellular during the hepatocyte phase; recent reports sug- GBCAs, which indicates short injection duration gests a 309angle.30,31 A greater ‰ip angle can be es- when using same injection rate. Since short injec- pecially helpful when imaging patients with chronic tion duration narrows optimal time window for ar- liver disease, whose impaired hepatic function terial phase, it is necessary to adjust individual fac- and/or gadoxetic acid uptake may make it di‹cult tors that aŠect the arrival time of the contrast agent to obtain su‹cient tumor-to-liver contrast.32 Respi- to the aorta. Bolus timing techniques, such as the ratory-triggered high resolution 3D-GRE sequence use of automated bolus detection algorithms or the using navigator echo has been proposed to obtain ‰uoroscopic triggering method, are particularly hepatocyte phase images with improved spatial

Vol. 12 No. 3, 2013 164 U. Motosugi et al. resolution, especially for the z axis.33 Coronal or decrease the signal intensity of the liver parenchy- sagittal images of good quality can be reconstruct- ma, which, in turn, may aŠect ADC measurement ed from the isovoxel axial images with submilli- in the liver parenchyma.39 meter spatial resolution.33,34 Roles of Gadoxetic Acid in Liver MR Imaging How to shorten examination time A 20-min waiting period for the hepatocyte Improving sensitivity of small hepatic metastasis phase is su‹cient to obtain other sequences, so detection many researchers have attempted to shorten exami- A meta-analysis published in 2010 suggested the nation time by obtaining T2-weighted and diŠu- superiority of contrast-enhanced MR imaging to sion-weighted images after gadoxetic acid adminis- contrast-enhanced CT for facilitating diagnosis of 45,46 tration. Acquisition of T2-weighted images after hepatic metastasis from colorectal carcinoma. gadoxetic acid injection has been demonstrated In general, there is no arguing that the high con- without compromise of image contrast. Gadoxetic trast resolution of MR imaging improves diagnoses 47 acid has both T2*-andT1-shortening eŠects. The of small focal liver lesions (Fig. 2a, b). However, T2*-shortening eŠect does not aŠect the image con- no publication shows the better utility of gadoxetic trast of the liver on fast spin-echo T2-weighted fast acid than superparamagnetic iron oxide (SPIO) for spin-echo images and does not decrease tumor con- detecting small liver lesions.48,49 Combined with dif- 35–39 spicuity or sensitivity on T2-weighted images. fusion-weighted imaging, gadoxetic acid-enhanced This is not true, though, for the bile duct, in MR imaging can yield better diagnostic accuracy which a much higher concentration of excreted and sensitivity than its use alone.50,51 The superio- gadoxetic acid than in the liver parenchyma leads to rity of gadoxetic acid-enhanced MR imaging to con- a signiˆcant decrease in bile duct signals and degra- trast-enhanced CT in patients with liver metasta- dation of T2-weighted MR cholangiopancreatog- ses from pancreatic carcinoma has also been ar- raphy (MRCP).40,41 Kim and colleagues, however, gued.52 Considering the added value of MRCP,53 reported that post-contrast MRCP is acceptable if gadoxetic acid-enhanced MR imaging can be a ˆrst- obtained within a few minutes of gadoxetic acid in- line examination for patients with suspected pan- jection42 because excretion of the contrast agent creatic carcinoma. into the bile duct does not start until approximately 5 to 10 min after injection.43 Lesion characterization DiŠusion-weighted images can also be obtained Hemangioma versus hepatic metastasis after gadoxetic acid administration. The presence Hemangiomas are liver tumors that, when small, of gadoxetic acid in the liver parenchyma does not are most frequently confused with metastatic liver aŠect tumor conspicuity or measurement of the ap- tumors. Traditional MR imaging ˆndings, such as parent diŠusion coe‹cient (ADC) of tumors on brightness on T2-weighted images, are essential for diŠusion-weighted images.35,38,44 Nevertheless, the distinguishing small hemangiomas from metastasis. T2*-shortening eŠect of the contrast agent can On hepatocyte phase images, hemangiomas display

Fig. 2. A case of liver metastases from rectal carcinoma. Metastasis in the posterior segment (arrow) is well visualized in contrast-enhanced computed tomography (CT) (a) and gadoxetic acid-enhanced hepatocyte phase image (b). Another small metastasis is only detected in hepatocyte phase image (b,arrow head).

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 165 a clear, low intensity signal because they contain no diagnosis of hemangioma. When comparing an hepatocytes that take up contrast agent. Thus, it is ability of tumor characterization between heman- impossible to observe the most characteristic ˆnd- gioma and metastasis, conventional GBCAs has ing of hemangioma, delayed enhancement or pool- advantage over gadoxetic acid.57 However, the ex- ing of the contrast agent, on delayed phase or hepa- cellent sensitivity of MR imaging with gadoxetic tocyte phase images. acid in detecting small lesions more than makes up We suppose that gadoxetic acid has no added for the disadvantages if multiparametric MR imag- value for evaluating hemangioma. Thus, 2 main ing evaluation is done. ˆndings, already known as useful features in conventional dynamic CT or MR imaging, can be used Distinguishing between focal nodular hyperplasia to distinguish hemangioma from metastasis when (FNH) and hepatic adenoma (HA) gadoxetic acid-enhanced MR imaging is used-ring- FNH and HA are focal lesions that develop in like enhancement on arterial phase images, which is the liver in the absence of chronic liver disease. observed more frequently in metastasis than Both types of lesion have been linked to use of oral hemangioma54 (Fig. 3a), and very high signal inten- contraceptives and commonly observed in the livers 55 sity on T2-weighted images, which is more com- of young women, and the lesions are characterized moninhemangiomathanmetastasis(Fig.3b).3 Be- by the presence of hypervascularity on arterial cause they contain similar concentrations of con- phase images. FNH is believed to be a regenerative trast agent, hemangiomas and vessels exhibit simi- hyperplastic response of hepatocytes secondary to lar signal intensities during the dynamic phases. localized vascular abnormalities, is usually man- Further, dot-like enhancements on arterial phase aged conservatively, and does not generally require images (Fig. 3c)3,54,56 and geographic or heteroge- resection. HA is a benign neoplasm composed of neous enhancement of larger hemangiomas on late hepatocytes and is treated with surgical resection if phase images (Fig. 3d)3 canalsofacilitateconˆdent the HA is considered at high risk for spontaneous

Fig. 3. Two major features for distinguishing liver metastasis from hemangioma are ring-like enhancement on arterial phase often observed in liver metasta-

sis (a) and bright, homogeneous very high intensity on T2-weighted image seen in hemangioma (b). Dot-like enhancement on arterial phase (c) and geographic moderate enhancement on late phase (d) are also frequently observed in hemangioma. Note: when comparing an ability of tumor characterization between hemangioma and metastasis gadoxetic acid has no added value to conventional extracellular gadolinium-based contrast agents.

Vol. 12 No. 3, 2013 166 U. Motosugi et al. hemorrhage or malignant transformation.30,32,39,58 and, in fact, may sometimes be more reliable than Recent studies have provided evidence that dynamic contrast-enhanced 16- to 64-slice mul- gadoxetic acid-enhanced MR imaging in general tidetector-row CT for this purpose.9,36,66,67 When and hepatocyte phase images in particular greatly using an extracellular GBCA, sensitivity and speci- facilitate the diŠerential diagnosis of these 2 le- ˆcity of MR imaging are excellent even for small sions. Most FNHs are strongly enhanced during the HCCs.68 However, some researchers suggest that arterial phase and show hyperintensity on hepato- the hepatocyte phase of gadoxetic acid might have cyte phase images (Fig. 4a, b).59,60 On the other added value to detect more HCCs than dynamic hand, hepatocyte phase images of HAs generally phases especially for less-experienced observers.69 show hypointensity. A central scar, an anomalous This is mainly the result of the conspicuity of portal area with abundant loose connective tissue, hypointense HCCs on hepatocyte phase images.70 is a well known and speciˆc imaging ˆnding of The decreased uptake of gadoxetic acid, or low in- FNH but not observed in all cases. However, we tensity on hepatocyte phase images, is easy to iden- should be aware that such scars appear on hepato- tify even if the lesion size is small, whereas washout cyte phase images with much lower intensity than on delayed dynamic phase CT is sometimes ambig- the remainder of the tumor.60 Although hepatocyte uous. However, approximately 10z of HCCs take phase ˆndings can aid diŠerential diagnosis, it is up gadoxetic acid and show iso- to hyperinten- also important to note that 10z of FNHs show sity.16 This suggests that the diagnosis of HCC hypointensity on hepatocyte phase images despite should not be based solely on dynamic and hepato- the fact they are non-neoplastic and benign.59,61 cyte phase images, but should also utilize conven- The central hypointense core of FNHs is generally tional sequences, such as T2-weighted imaging (Fig. larger in extent than the central scar observed in dy- 5a, b). namic phases.60,62 Even in such cases, however, the Hypervascular pseudolesions, also known as ar- peripheral hyperintense rim on hepatocyte phase terial-portal shunts, typically have a rectangular or might imply an uptake ability of the lesion, i.e., wedge shape, are commonly found at the periphery FNH rather than HA (Fig. 4b).60 of cirrhotic livers, and may mimic small hypervas- HAs have recently been subclassiˆed into 4 cate- cular HCCs.71 Even on extracellular GBCA-en- gories on the basis of genetic and pathologic fea- hanced MR imaging, most hypervascular pseudole- tures,63 but only a little has been studied on the role sions are easily recognized as benign, though some of gadoxetic acid for this subclassiˆcation.61 have a rounder shape and can be misdiagnosed as small HCCs. Uptake of gadoxetic acid, or isointen- Added value for hepatocellular carcinoma (HCC) sity to the surrounding liver parenchyma on hepa- Improving diagnosis of conventional HCCs tocyte phase images, is a key feature for conˆrming Arterial hypervascularity and subsequent wash- a hypervascular focus is benign72 (Fig. 6a, b). out are the widely recognized and accepted criteria Although hypervascular pseudolesions can some- for imaging diagnosis of HCCs.64,65 Several studies times show slightly lower intensity on the hepato- have shown that gadoxetic acid-enhanced MR imag- cyte phase, the diŠerence in signal between the le- ing is useful in diagnosing conventional HCCs sion and surrounding liver parenchyma is small

Fig. 4. A case of focal nodular hyperplasia (FNH). Hypervascularity on arterial phase (a, arrow) and subsequent uptake of gadoxetic acid on hepatocyte phase (b) are the characteristic features of FNH. Note, part of the non-enhanced area is seen within the lesion (b,arrow).

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 167

Fig. 5. A case of type C cirrhosis with hypervascular hepatocellular carcinoma (HCC). Arterial enhancement is observed in the lesion (a). A part of the tumor uptakes gadoxetic acid and shows as high signal intensity on hepatocyte phase image as the liver parenchyma (b). This atypical uptake of gadoxetic acid in the HCC is not uncommon.

Fig. 6. Hypervascular pseudolesion or arterial-portal shunt in cirrhotic liver (a, b). Round hypervascular focus is observed on arterial phase (a), which uptake gadoxetic acid as well as the surrounding liver parenchyma (b). The ˆnd- ings suggest the lesion should be pseudolesion rather than hepatocellular carcinoma (HCC).

Fig. 7. Small hypervascular hepatocellular carcinoma (HCC) in cirrhotic liver. Faint hypervascular nodule is detected in the lateral segment on arterial phase image (a). The nodule shows clear hypointensity on hepatocyte phase (b)and hyperintensity on diŠusion-weighted image (c).

Vol. 12 No. 3, 2013 168 U. Motosugi et al. despite the reduced level of gadoxetic acid uptake Advantage of assessing multi-step hepatocarcino- in the pseudolesion.73 Very low signal intensity of genesis: assessment of hypovascular HCCs hypervascular lesions compared with the surround- Although hypervascularity is an established crite- ingliverparenchyma,aswellashyperintensityon rion for diagnosing conventional HCCs, HCCs diŠusion-weighted images, can be reliable features may also show hypovascularity on arterial phase for diagnosing small HCC (Fig. 7a–c). images, especially in their early stages. It is widely Small hemangiomas may also mimic hypervascu- believed that HCC can develop in cirrhotic livers lar HCCs, and, as mentioned, hypointensity of he- from premalignant lesions or dysplastic nodules via mangiomas on hepatocyte phase images can some- multi-step carcinogenesis.74–76 During this process, times be misleading. Bright signals on T2-weighted a pathological feature called early HCC is clearly images imply non-solid pathology and should al- distinguished from conventional hypervascular ways be checked during the assessment of small HCC77,78 (Fig. 8a–e). For example, early HCCs are hypervascular lesions. usually smaller than 2 cm in size, and macroscopic pathological observations reveal a vaguely nodular

Fig. 8. AcaseoftypeCchronichepatitiswithconventional small hepatocellular carcinoma (HCC) (a, b)andearlyHCC(c, d). Hypervascular nodule on arterial phase (a) shows clear hypointensity on hepatocyte phase image (b). This nodule is pathologically diagnosed as moderately diŠerentiated HCC. Another small nodule without hypervascularity (c) also shows hypointensity on hepatocyte phase image (d). This nodule was surgically removed and diagnosed as early HCC. Macroscopically, the former nodule (moderately diŠerentiated HCC) is whitish with a well deˆned margin (e, arrow), whereas the latter nodule (early HCC) is vaguely nodular (e, arrowhead).

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 169 appearance and the absence of a pseudocapsule when assessing a cirrhotic nodule.86 surrounding the lesion. Further, early HCCs, unlike conventional HCCs, are not hypervascular, Room for Discussion though focalhypervascularity may beobserved within a lesion.77 Finally, early HCCs can show wash- Management of hypovascular hepatocellular nod- out on delayed-phase dynamic CT or decreased ules showing low intensity on hepatocyte phase im- portal perfusion on CT during arterial portogra- ages phy.79 The bulk of the literature suggests that hypovas- Though the use of conventional imaging modali- cular nodules with low intensity on hepatocyte ties to diagnose early HCC has been challenging, phase images can be malignant or at least premalig- the use of gadoxetic acid-enhanced MR imaging nant.80,81 If left untreated, more than 10z of these may at least partly alleviate this problem. Speciˆ- lesions will become hypervascular or conventional cally, early HCCs may appear on hepatocyte phase HCCs within a year. Among early HCCs identiˆed images as clearly hypointense, hypovascular lesions by gadoxetic acid-enhanced MR imaging, an espe- in cirrhotic livers.80 Gadoxetic acid-enhanced hepa- cially important risk factor for hypervasculariza- tocyte phase images have been reported to have the tion is size exceeding 10 mm87,88 (Fig. 9a–d). How- highest sensitivity and speciˆcity for diagnosing ever, this does not mean that this type of lesion, early HCCs.81 After reviewing cirrhotic liver nod- such as one that is hypovascular and hypointense ules that showed atypical enhancement patterns on during hepatocyte phase imaging, should always be dynamic contrast-enhanced studies, Golˆeri and treated as a hypervascular HCC. Future work, in- associates also concluded that hypointensity on cluding prospective randomized studies, are re- hepatocyte phase images is the strongest marker of quired to determine whether it is appropriate malignancy.82 Although some investigators have and/or necessary to treat these early stage HCCs. suggested that dysplastic nodules can also be seen as hypointense nodules on hepatocyte phase im- HCCs taking up gadoxetic acid ages,83–85 we believe that low intensity on hepato- As discussed, low signal intensity on hepatocyte cyte phase is the key feature suggesting malignancy phase images can indicate the malignant transpor-

Fig. 9. Small hypointense nodule on hepatocyte phase in cirrhotic liver (a,arrow), which does not show hypervascularity on arterial phase (b). Hepatocyte phase image after 6 months of follow-up shows increased size (c). Hypervas- cularity in the enlarged nodule (d) suggests transformation from early hepatocellular carcinoma (HCC) to conventional hypervascular HCC.

Vol. 12 No. 3, 2013 170 U. Motosugi et al. tation of hepatocellular lesions.80–82 However, high hypervascular HCCs, some hypovascular HCCs intensity on hepatocyte phase does not always sug- can show high signal intensity, or increased uptake gest benignity. Approximately 10 to 20z of con- of gadoxetic acid, in the hepatocyte phase. Hypo- ventional HCCs can show high intensity or take up vascular hepatocellular nodules showing high in- gadoxetic acid during hepatocyte phase imaging.89 tensity during hepatocyte phase imaging can be ear- There have been reports that OATP-8, an anion ly HCC92 (Fig. 10a, b) and sometimes progress to transporter on the hepatocyte surface, is overex- hypervascular HCCs after several months (Fig. pressed in these lesions.90,91 One study suggested 11). Kobayashi and colleagues reported gadoxetic that none of these lesions was poorly diŠerentiated acid-enhanced MR imaging ˆndings of high grade HCC.83 ThismayindicatethatlossofOATP-8ex- hepatocellular lesions, or so-called early HCCs, de- pression, or loss of signal intensity on the hepato- tected by angiography-assisted CT that revealed cyte phase, may not happen until well after malig- decreased portal ‰ow to the lesions during arterial nant transformation. It would be quite interesting portography and a small hypervascular focus with- if the hyperintensity of HCC is associated with in the tumor during hepatic arteriography. The tumor recurrence rate or patient prognosis. authors reported that approximately 20z of these The existence of high intensity nodules in hepato- lesions showed isointensity on gadoxetic acid-en- cyte phase images can be a source of confusion for hanced hepatocyte phase images,93 which suggests characterizing HCCs by gadoxetic acid-enhanced that it might be even more di‹cult than expected to MR imaging. However, it is usually easy to diag- detect early HCCs in cirrhotic livers using only nose HCC even if the tumor shows high signal in- gadoxetic acid-enhanced MR images. Further stud- tensities relative to the liver parenchyma because ies, including those incorporating longitudinal ob- other sequence images show additional characteris- servations, are necessary to address this issue. tic features, such as mildly high intensity on T2- It is important to remember that the signal inten- weighted images, high signal intensity on diŠusion- sity of hepatocyte phase images varies depending weighted images, heterogeneous hypervascularity on hepatic function, degree of ˆbrosis, and portal on the arterial phase (unlike FNH), and heter- ‰ow. Thus, the relative intensity of lesions and liver ogeneous mass-like signal intensities on subsequent parenchyma is not always identical among patients. washout during late phase imaging. Whenever hepatocyte phase intensity is discussed, it is vital to conˆrm enhancement by comparing sig- Hypovascular hepatocellular nodules showing high nals of the liver with those of extracellular com- intensity on hepatocyte-phase imaging partments, such as the spleen. The above discussion raises the question of whether hepatocellular lesions showing high inten- Quantitative assessment of diŠuse liver disease sity in the hepatocyte phase are always benign, and Because the uptake of gadoxetic acid into the the answer is no. As is the case for conventional liver parenchyma re‰ects hepatocyte function,

Fig. 10. Early hepatocellular carcinoma (HCC) taking up gadoxetic acid. Hepatocyte phase image shows small nodule at the posterior edge of the liver (arrow) with high signal intensity (a). No hypervascularity is observed on arterial phase image (not shown). The pathological diagnosis is early HCC. Tumor cells in the top of the microscopic image show cellular atypia and increased cel- lularity (b). This is the same patient as ``Case 2'' reported by Chen and associates.92

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 171

Fig. 11. A case of type C cirrhosis with hypervascular hepatocellular carcinoma (HCC) on arterial phase (a, arrowhead) and hypointensity on hepatocyte phase (b, arrowhead). Post-ablation changes are seen on images obtained 6 months after therapy (c, d, arrowhead). Another nodule without hypervascularity on arterial phase (a, arrow) shows high intensity on hepatocyte phase (b, arrow). The nodule became hypervascular during the 6 months of follow-up (c, arrow). The signal drop on hepatocyte phase also appears (d, arrow).

many researchers attempt to use gadoxetic acid- patients with cirrhosis, diagnosing hypovascular enhanced MR imaging, and especially hepatocyte HCCs, and facilitating diŠerential diagnosis be- phase imaging, to make non-morphological or tween FNH and HA. As a result, this contrast a- functional assessments of the liver.94 For example, gent has changed strategies for managing patients because the liver enhancement ratio can, with some with liver disease, and clinical guidelines will soon correction, be correlated with degree of ˆbrosis, be revised to re‰ect this new trend in liver MR this technique has been used to determine stage of imaging. Although gadoxetic acid has many advan- liver ˆbrosis.95–97 The ratio of hepatocyte phase en- tages, radiologists should be aware of its potential hancement has also been used to assess local hepat- drawbacks and always be cautious when employing ic function.98 Thus, non-morphological or quan- this product in clinical settings. titative MR imaging using gadoxetic acid as well as quantitative assessment using diŠusion-weighted References and perfusion imaging may represent the new generation of projects.86 For these techniques to be 1. Frydrychowicz A, Lubner MG, Brown JJ, et al. useful, it will be necessary to evaluate their re- Hepatobiliary MR imaging with gadolinium-based peatability and reproducibility and compare results contrast agents. J Magn Reson Imaging 2012; 35: from these methods with those produced by pre-ex- 492–511. isting techniques, such as laboratory analyses.99 2. Doo KW, Lee CH, Choi JW, Lee J, Kim KA, Park CM. ``Pseudo washout'' sign in high-‰ow hepatic hemangioma on gadoxetic acid contrast-enhanced Conclusions MRI mimicking hypervascular tumor. AJR Am J Roentgenol 2009; 193:W490–W496. Compared to conventional extracellular gadolin- 3. Motosugi U, Ichikawa T, Onohara K, et al. Distin- ium-based contrast agents, gadoxetic acid is better guishing hepatic metastasis from hemangioma us- for detecting hepatic metastases, distinguishing ing gadoxetic acid-enhanced magnetic resonance hypervascular pseudolesions from small HCCs in

Vol. 12 No. 3, 2013 172172 U. Motosugi et al.

imaging. Invest Radiol 2011; 46:359–365. 1986; 147:1283–1287. 4. Motosugi U, Ichikawa T, Sou H, et al. Liver 16. Oshio K, Feinberg DA. GRASE (Gradient- and parenchymal enhancement of hepatocyte-phase spin-echo) imaging: a novel fast MRI technique. images in Gd-EOB-DTPA-enhanced MR imaging: Magn Reson Med 1991; 20:344–349. which biological markers of the liver function 17. Haradome H, Grazioli L, Tsunoo M, et al. Can aŠect the enhancement? J Magn Reson Imaging MR ‰uoroscopic triggering technique and slow rate 2009; 30:1042–1046. injection provide appropriate arterial phase images 5. Kuwatsuru R, Brasch RC, Muhlerä A, et al. Deˆni- with reducing artifacts on gadoxetic acid-DTPA tion of liver tumors in the presence of diŠuse liver (Gd-EOB-DTPA)-enhanced hepatic MR imaging? disease: comparison of ˆndings at MR imaging J Magn Reson Imaging 2010; 32:334–340. with positive and negative contrast agents. Radiol- 18. Motosugi U, Ichikawa T, Sou H, et al. Dilution ogy 1997; 202:131–138. method of gadolinium ethoxybenzyl diethylenetri- 6. Tamada T, Ito K, Higaki A, et al. Gd-EOB- aminepentaacetic acid (Gd-EOB-DTPA)-enhanced DTPA-enhanced MR imaging: evaluation of he- magnetic resonance imaging (MRI). J Magn Reson patic enhancement eŠects in normal and cirrhotic Imaging 2009; 30:849–854. livers. Eur J Radiol 2011; 80:e311–e316. 19. Tamada T, Ito K, Yoshida K, et al. Comparison of 7. Higaki A, Tamada T, Sone T, et al. Potential clini- three diŠerent injection methods for arterial phase cal factors aŠecting hepatobiliary enhancement at of Gd-EOB-DTPA enhanced MR imaging of the Gd-EOB-DTPA-enhanced MR imaging. Magn Res- liver. Eur J Radiol 2011; 80:e284–e288. on Imaging 2012; 30:689–693. 20. Golˆeri R, Renzulli M, Lucidi V, Corcioni B, 8. Motosugi U, Ichikawa T, Muhi A, et al. Magnetic Trevisani F, Bolondi L. Contribution of the hepa- resonance elastography as a predictor of insuf- tobiliary phase of Gd-EOB-DTPA-enhanced MRI ˆcient liver enhancement on gadoxetic acid-en- to dynamic MRI in the detection of hypovascular hanced hepatocyte-phase magnetic resonance imag- small (Ã2 cm) HCC in cirrhosis. Eur Radiol 2011; ing in patients with type C hepatitis and Child- 21:1233–1242. Pugh class A disease. Invest Radiol 2012; 47:566– 21.OnishiH,KimT,ImaiY,etal.Hypervascular 570. hepatocellular carcinomas: detection with gadox- 9. Tamada T, Ito K, Sone T, et al. Dynamic contrast- etate disodium-enhanced MR imaging and mul- enhanced magnetic resonance imaging of abdomi- tiphasic multidetector CT. Eur Radiol 2012; 22: nal solid organ and major vessel: comparison of 845–854. enhancement eŠect between Gd-EOB-DTPA and 22. Motosugi U, Ichikawa T, Sano K, et al. Double- Gd-DTPA. J Magn Reson Imaging 2009; 29:636– dose gadoxetic acid-enhanced magnetic resonance 640. imaging in patients with chronic liver disease. In- 10. Kuhn ä JP, Hegenscheid K, Siegmund W, Froehlich vest Radiol 2011; 46:141–145. CP, Hosten N, Puls R. Normal dynamic MRI en- 23. Reimer P, Rummeny EJ, Shamsi K, et al. Phase II hancement patterns of the upper abdominal or- clinical evaluation of Gd-EOB-DTPA: dose, safety gans: gadoxetic acid compared with gadobutrol. aspects, and pulse sequence. Radiology 1996; 199: AJR Am J Roentgenol 2009; 193:1318–1323. 177–183. 11. Zech CJ, Vos B, Nordell A, et al. Vascular en- 24. Huppertz A, Haraida S, Kraus A, et al. Enhance- hancement in early dynamic liver MR imaging in ment of focal liver lesions at gadoxetic acid-en- an animal model: comparison of two injection reg- hanced MR imaging: correlation with histopatho- imen and two diŠerent doses Gd-EOB-DTPA logic ˆndings and spiral CT–initial observations. (gadoxetic acid) with standard Gd-DTPA. Invest Radiology 2005; 234:468–478. Radiol 2009; 44:305–310. 25. Yamamoto A, Tamada T, Sone T, Higashi H, 12. Hamm B, Staks T, Muhlerä A, et al. Phase I clinical Yamashita T, Ito K. Gd-EOB-DTPA-enhanced evaluation of Gd-EOB-DTPA as a hepatobiliary magnetic resonance imaging ˆndings of nondiŠuse MR contrast agent: safety, pharmacokinetics, and fatty change of the liver. J Comput Assist Tomogr MR imaging. Radiology 1995; 195:785–792. 2010; 34:868–873. 13. Tanimoto A, Lee JM, Murakami T, Huppertz A, 26. Tamada T, Ito K, Sone T, Kanki A, Sato T, Kudo M, Grazioli L. Consensus report of the 2nd Higashi H. Gd-EOB-DTPA enhanced MR imag- International Forum for Liver MRI. Eur Radiol ing: evaluation of biliary and renal excretion in 2009; 19 Suppl 5:S975–S989. normal and cirrhotic livers. Eur J Radiol 2011; 80: 14. Tanimoto A, Higuchi N, Ueno A. Reduction of e207–e211. ringing artifacts in the arterial phase of gadoxetic 27. Motosugi U, Ichikawa T, Tominaga L, et al. Delay acid-enhanced dynamic MR imaging. Magn Reson before the hepatocyte phase of Gd-EOB-DTPA- Med Sci 2012; 11:91–97. enhanced MR imaging: is it possible to shorten the 15. Lufkin RB, Pusey E, Stark DD, Brown R, Leikind examination time? Eur Radiol 2009; 19:2623–2629. B, Hanafee WN. Boundary artifact due to trunca- 28. Jeong HT, Kim MJ, Park MS, et al. Detection of tion errors in MR imaging. AJR Am J Roentgenol liver metastases using gadoxetic-enhanced dynamic

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 173173

and 10- and 20-minute delayed phase MR imaging. 40. Ringe KI, Gupta RT, Brady CM, et al. Respira-

J Magn Reson Imaging 2012; 35:635–643. tory-triggered three-dimensional T2-weighted MR 29. Sofue K, Tsurusaki M, Tokue H, Arai Y, cholangiography after injection of gadoxetate dis- Sugimura K. Gd-EOB-DTPA-enhanced 3.0T MR odium: is it still reliable? Radiology 2010; 255:451– imaging: quantitative and qualitative comparison 458. of hepatocyte-phase images obtained 10 min and 41. Nakamura Y, Ohmoto T, Saito T, Kajima T, 20 min after injection for the detection of liver Nishimaru E, Ito K. EŠects of gadolinium-ethox-

metastases from colorectal carcinoma. Eur Radiol ybenzyl-diethylenetriamine pentaacetic acid on T2- 2011; 21:2336–2343. weighted MRCP. Magn Reson Med Sci 2009; 8: 30. Bashir MR, Merkle EM. Improved liver lesion con- 143–148.

spicuity by increasing the ‰ip angle during hepato- 42. KimKA,KimMJ,ParkMS,etal.OptimalT2- cyte phase MR imaging. Eur Radiol 2011; 21:291– weighted MR cholangiopancreatographic images 294. can be obtained after administration of gadoxetic 31. Bashir MR, Husarik DB, Ziemlewicz TJ, Gupta acid. Radiology 2010; 256:475–484. RT, Boll DT, Merkle EM. Liver MRI in the 43. Ringe KI, Husarik DB, Gupta RT, Boll DT, Mer- hepatocyte phase with gadolinium-EOB-DTPA: kle EM. Hepatobiliary transit times of gadoxetate does increasing the ‰ip angle improve conspicuity disodium (Primovist) for protocol optimization and detection rate of hypointense lesions? J Magn of comprehensive MR imaging of the biliary sys- Reson Imaging 2012; 35:611–616. tem–what is normal? Eur J Radiol 2011; 79:201– 32. Haradome H, Grazioli L, Al manea K, et al. 205. Gadoxetic acid disodium-enhanced hepatocyte 44. Shimofusa R, Ueda T, Kishimoto T, et al. Magnet- phase MRI: can increasing the ‰ip angle improve ic resonance imaging of hepatocellular carcinoma: focal liver lesion detection? J Magn Reson Imaging a pictorial review of novel insights into pathophysi- 2012; 35:132–139. ological features revealed by magnetic resonance 33. Nagle SK, Busse RF, Brau AC, et al. High resolu- imaging. J Hepatobiliary Pancreat Sci 2010; 17: tion navigated three-dimensional T(1)-weighted 583–589. hepatobiliary MRI using gadoxetic acid optimized 45. Floriani I, Torri V, Rulli E, et al. Performance of for 1.5 Tesla. J Magn Reson Imaging 2012; 36: imaging modalities in diagnosis of liver metastases 890–899. from colorectal cancer: a systematic review and 34. Frydrychowicz A, Nagle SK, D'Souza SL, Vigen meta-analysis. J Magn Reson Imaging 2010; 31: KK, Reeder SB. Optimized high-resolution con- 19–31. trast-enhanced hepatobiliary imaging at 3 tesla: a 46. Niekel MC, Bipat S, Stoker J. Diagnostic imaging cross-over comparison of gadobenate dimeglumine of colorectal liver metastases with CT, MR imag- and gadoxetic acid. J Magn Reson Imaging 2011; ing, FDG PET, and/or FDG PET/CT: a meta- 34:585–594. analysis of prospective studies including patients 35. Saito K, Araki Y, Park J, et al. EŠect of Gd-EOB- who have not previously undergone treatment.

DTPA on T2-weighted and diŠusion-weighted im- Radiology 2010; 257:674–684. ages for the diagnosis of hepatocellular carcinoma. 47. Huppertz A, Balzer T, Blakeborough A, et al. Im- J Magn Reson Imaging 2010; 32:229–234. proved detection of focal liver lesions at MR imag-

36. Tamada T, Ito K, Yoshida K, et al. T2-weighted ing: multicenter comparison of gadoxetic acid-en- magnetic resonance imaging of the liver: evalua- hanced MR images with intraoperative ˆndings. tion of the eŠect in signal intensity after Gd-EOB- Radiology 2004; 230:266–275. DTPA enhancement. J Comput Assist Tomogr 48. Muhi A, Ichikawa T, Motosugi U, et al. Diagnosis 2010; 34:182–186. of colorectal hepatic metastases: comparison of 37. Kim YK, Kwak HS, Kim CS, Han YM. Detection contrast-enhanced CT, contrast-enhanced US, su- and characterization of focal hepatic tumors: a perparamagnetic iron oxide-enhanced MRI, and

comparison of T2-weighted MR images before and gadoxetic acid-enhanced MRI. J Magn Reson after the administration of gadoxectic acid. J Imaging 2011; 34:326–335. Magn Reson Imaging 2009; 30:437–443. 49. Kim YK, Lee YH, Kwak HS, Kim CS, Han YM. 38. Choi JS, Kim MJ, Choi JY, Park MS, Lim JS, Detection of liver metastases: gadoxetic acid-en- Kim KW. DiŠusion-weighted MR imaging of liver hanced three-dimensional MR imaging versus on 3.0-Tesla system: eŠect of intravenous adminis- ferucarbotran-enhanced MR imaging. Eur J Radi- tration of gadoxetic acid disodium. Eur Radiol ol 2010; 73:131–136. 2010; 20:1052–1060. 50. Kim YK, Lee MW, Lee WJ, et al. Diagnostic ac- 39. Muhi A, Ichikawa T, Motosugi U, Sou H, Sano K, curacy and sensitivity of diŠusion-weighted and of

Araki T. DiŠusion- and T2-weighted MR imaging gadoxetic acid-enhanced 3-T MR imaging alone or of the liver: eŠect of intravenous administration of in combination in the detection of small liver gadoxetic acid disodium. Magn Reson Med Sci metastasis (Ã1.5 cm in diameter). Invest Radiol 2012; 11:185–191. 2012; 47:159–166.

Vol. 12 No. 3, 2013 174174 U. Motosugi et al.

51. Kanematsu M, Goshima S, Watanabe H, et al. date. Radiographics 2011; 31:1529–1543. DiŠusion/perfusion MR imaging of the liver: prac- 64. Bruix J, Sherman M; American Association for the tice, challenges, and future. Magn Reson Med Sci Study of Liver Diseases. Management of hepato- 2012; 11:151–161. cellular carcinoma: an update. Hepatology 2011; 52. Motosugi U, Ichikawa T, Morisaka H, et al. De- 53:1020–1022. tection of pancreatic carcinoma and liver metas- 65. Study Group of Evidence-based Practice Guideline tases with gadoxetic acid-enhanced MR imaging: for Liver Cancer. Evidence-based Liver Cancer comparison with contrast-enhanced multi-detector Practice Guideline. Tokyo: KANEHARA & CO., row CT. Radiology 2011; 260:446–453. LTD., 2009. 53. Park HS, Lee JM, Choi HK, Hong SH, Han JK, 66. Di Martino M, Marin D, Guerrisi A, et al. Intrain- Choi BI. Preoperative evaluation of pancreatic dividual comparison of gadoxetate disodium-en- cancer: comparison of gadolinium-enhanced dy- hanced MR imaging and 64-section multidetector namic MRI with MR cholangiopancreatography CT in the detection of hepatocellular carcinoma in versus MDCT. J Magn Reson Imaging 2009; 30: patients with cirrhosis. Radiology 2010; 256:806– 586–595. 816. 54. Mitchell DG, Saini S, Weinreb J, et al. Hepatic 67. Murakami T, Okada M, Hyodo T. CT versus MR metastases and cavernous hemangiomas: distinc- imaging of hepatocellular carcinoma: toward im- tion with standard- and triple-dose gadoteridol-en- proved treatment decisions. Magn Reson Med Sci hanced MR imaging. Radiology 1994; 193:49–57. 2012; 11:75–81. 55. Soyer P, Dufresne AC, Somveille E, Scherrer A. 68. Karadeniz-Bilgili MY, Braga L, Birchard KR, et al.

Hepatic cavernous hemangioma: appearance on T2 Hepatocellular carcinoma missed on gadolinium -weighted fast spin-echo MR imaging with and enhanced MR imaging, discovered in liver ex- without fat suppression. AJR Am J Roentgenol plants: retrospective evaluation. J Magn Reson Im- 1997; 168:461–465. aging 2006; 23:210–215. 56. Imai Y, Murakami T, Hori M, et al. Hypervascu- 69. Ahn SS, Kim MJ, Lim JS, Hong HS, Chung YE, lar hepatocellular carcinoma: combined dynamic Choi JY. Added value of gadoxetic acid-enhanced MDCT and SPIO-enhanced MRI versus combined hepatobiliary phase MR imaging in the diagnosis CTHA and CTAP. Hepatol Res 2008; 38:147–158. of hepatocellular carcinoma. Radiology 2010; 255: 57. Vogl TJ, Kummelä S, Hammerstingl R, et al. Liver 459–466. tumors: comparison of MR imaging with Gd-EOB- 70. Frericks BB, Loddenkemper C, Huppertz A, et al. DTPA and Gd-DTPA. Radiology 1996; 200:59– Qualitative and quantitative evaluation of hepato- 67. cellular carcinoma and cirrhotic liver enhancement 58. Takara K, Saito K, Kusama H, et al. Gd-EOB- using Gd-EOB-DTPA. AJR Am J Roentgenol DTPA-enhanced MR imaging ˆndings of hepato- 2009; 193:1053–1060. cellular adenoma: correlation with pathological 71. Hwang SH, Yu JS, Kim KW, Kim JH, Chung JJ. ˆndings. Magn Reson Med Sci 2011; 10:245–249. Small hypervascular enhancing lesions on arterial 59. Zech CJ, Grazioli L, Breuer J, Reiser MF, Schoen- phase images of multiphase dynamic computed berg SO. Diagnostic performance and description tomography in cirrhotic liver: fate and implica- of morphological features of focal nodular hyper- tions. J Comput Assist Tomogr 2008; 32:39–45. plasia in Gd-EOB-DTPA-enhanced liver magnetic 72. Sun HY, Lee JM, Shin CI, et al. Gadoxetic acid- resonance imaging: results of a multicenter trial. enhanced magnetic resonance imaging for diŠeren- Invest Radiol 2008; 43:504–511. tiating small hepatocellular carcinomas (º or ＝2 60. Mohajer K, Frydrychowicz A, Robbins JB, cm in diameter) from arterial enhancing pseudole- LoeŒer AG, Reed TD, Reeder SB. Characteriza- sions: special emphasis on hepatobiliary phase tion of hepatic adenoma and focal nodular hyper- imaging. Invest Radiol 2010; 45:96–103. plasia with gadoxetic acid. J Magn Reson Imaging 73. Motosugi U, Ichikawa T, Sou H, et al. Distin- 2012; 36:686–696. guishing hypervascular pseudolesions of the liver 61. Grazioli L, Bondioni MP, Haradome H, et al. from hypervascular hepatocellular carcinomas Hepatocellular adenoma and focal nodular hyper- with gadoxetic acid-enhanced MR imaging. Radi- plasia: value of gadoxetic acid-enhanced MR imag- ology 2010; 256:151–158. ing in diŠerential diagnosis. Radiology 2012; 262: 74. Kudo M. Multistep human hepatocarcinogenesis: 520–529. correlation of imaging with pathology. J Gastroen- 62. Karam AR, Shankar S, Surapaneni P, Kim YH, terol 2009; 44 Suppl 19:112–118. Hussain S. Focal nodular hyperplasia: central scar 75. Matsui O. Imaging of multistep human hepatocar- enhancement pattern using gadoxetate disodium. J cinogenesis by CT during intra-arterial contrast in- Magn Reson Imaging 2010; 32:341–344. jection. Intervirology 2004; 47:271–276. 63. Katabathina VS, Menias CO, Shanbhogue AK, 76. Efremidis SC, Hytiroglou P. The multistep process Jagirdar J, Paspulati RM, Prasad SR. Genetics of hepatocarcinogenesis in cirrhosis with imaging and imaging of hepatocellular adenomas: 2011 up- correlation. Eur Radiol 2002; 12:753–764.

Magnetic Resonance in Medical Sciences Current Knowledge of Gadoxetic Acid-MRI 175175

77. International Consensus Group for Hepatocellular 89. Saito K, Kotake F, Ito N, et al. Gd-EOB-DTPA Neoplasia. Pathologic diagnosis of early hepato- enhanced MRI for hepatocellular carcinoma: cellular carcinoma: a report of the International quantitative evaluation of tumor enhancement in Consensus Group for Hepatocellular Neoplasia. hepatobiliary phase. Magn Reson Med Sci 2005; 4: Hepatology 2009; 49:658–664. 1–9. 78. Sakamoto M, Hirohashi S, Shimosato Y. Early 90. Kitao A, Zen Y, Matsui O, et al. Hepatocellular stages of multistep hepatocarcinogenesis: adenom- carcinoma: signal intensity at gadoxetic acid-en- atous hyperplasia and early hepatocellular carcino- hanced MR imaging–correlation with molecular ma. Hum Pathol 1991; 22:172–178. transporters and histopathologic features. Radiol- 79. Takayasu K, Muramatsu Y, Furukawa H, et al. ogy 2010; 256:817–826. Early hepatocellular carcinoma: appearance at CT 91. Tsuboyama T, Onishi H, Kim T, et al. Hepatocel- during arterial portography and CT arteriography lular carcinoma: hepatocyte-selective enhancement with pathologic correlation. Radiology 1995; 194: at gadoxetic acid-enhanced MR imaging–correla- 101–105. tion with expression of sinusoidal and canalicular 80. Kogita S, Imai Y, Okada M, et al. Gd-EOB- transporters and bile accumulation. Radiology DTPA-enhanced magnetic resonance images of 2010; 255:824–833. hepatocellular carcinoma: correlation with histo- 92. Chen N, Motosugi U, Sano K, et al. Early logical grading and portal blood ‰ow. Eur Radiol hepatocellular carcinomas showing isointensity or 2010; 20:2405–2413. hyperintensity in gadoxetic acid-enhanced, hepato- 81. Sano K, Ichikawa T, Motosugi U, et al. Imaging cyte-phase magnetic resonance images. J Comput study of early hepatocellular carcinoma: usefulness Assist Tomogr 2013 (in press). of gadoxetic acid-enhanced MR imaging. Radiolo- 93. Kobayashi S, Matsui O, Gabata T, et al. Gadolini- gy 2011; 261:834–844. um ethoxybenzyl diethylenetriamine pentaacetic 82. Golˆeri R, Grazioli L, Orlando E, et al. Which is acid-enhanced magnetic resonance imaging ˆnd- the best MRI marker of malignancy for atypical ings of borderline lesions at high risk for progres- cirrhotic nodules: hypointensity in hepatobiliary sion to hypervascular classic hepatocellular carci- phase alone or combined with other features? Clas- noma. J Comput Assist Tomogr 2011; 35:181–186. siˆcation after Gd-EOB-DTPA administration. J 94. Moriyama Y, Saito S, Kobayashi S, et al. Evalua- Magn Reson Imaging 2012; 36:647–658. tion of concanavalin A-induced acute liver injury 83. Kitao A, Matsui O, Yoneda N, et al. The uptake in rats using an empirical mathematical model and transporter OATP8 expression decreases during dynamic contrast-enhanced MR imaging with Gd- multistep hepatocarcinogenesis: correlation with EOB-DTPA. Magn Reson Med Sci 2012; 11:53– gadoxetic acid enhanced MR imaging. Eur Radiol 60. 2011; 21:2056–2066. 95. Watanabe H, Kanematsu M, Goshima S, et al. 84. Matsui O. Multistep hepatocarcinogenesis and ear- Staging hepatic ˆbrosis: comparison of gadoxetate ly hepatocellular carcinoma: imaging-pathological disodium-enhanced and diŠusion-weighted MR correlation. In: Arii S, Matsui O, eds. Early diag- imaging–preliminaryobservations.Radiology2011; nosis of hepatocellular carcinoma: imaging and 259:142–150. molecular marker. Tokyo: Arc Medium, 2012; 15– 96. Kanki A, Tamada T, Higaki A, et al. Hepatic 34. parenchymal enhancement at Gd-EOB-DTPA-en- 85. Bartolozzi C, Battaglia V, Bargellini I, et al. Con- hanced MR imaging: correlation with morphologi- trast-enhanced magnetic resonance imaging of 102 cal grading of severity in cirrhosis and chronic nodules in cirrhosis: correlation with histological hepatitis. Magn Reson Imaging 2012; 30:356–360. ˆndings on explanted livers. Abdom Imaging 2013; 97. Chen BB, Hsu CY, Yu CW, et al. Clinical and 38:290–296. histologic implications of delayed hepatobiliary en- 86. Van Beers BE, Pastor CM, Hussain HK. hancement on magnetic resonance imaging with Primovist, Eovist: what to expect? J Hepatol 2012; gadolinium ethoxybenzyl diethylenetriaminepenta- 57:421–429. acetic acid. Invest Radiol 2012; 47:649–655. 87. Motosugi U, Ichikawa T, Sano K, et al. Outcome 98. Yamada A, Hara T, Li F, et al. Quantitative evalu- of hypovascular hepatic nodules revealing no ation of liver function with use of gadoxetate dis- gadoxetic acid uptake in patients with chronic liver odium-enhanced MR imaging. Radiology 2011; disease. J Magn Reson Imaging 2011; 34:88–94. 260:727–733. 88. Kumada T, Toyoda H, Tada T, et al. Evolution of 99. Motosugi U, Ichikawa T, Oguri M, et al. Staging hypointense hepatocellular nodules observed only liver ˆbrosis by using liver-enhancement ratio of in the hepatobiliary phase of gadoxetate disodium- gadoxetic acid-enhanced MR imaging: comparison enhanced MRI. AJR Am J Roentgenol 2011; 197: with aspartate aminotransferase-to-platelet ratio 58–63. index. Magn Reson Imaging 2011; 29:1047–1052.

Vol. 12 No. 3, 2013