Consensus Statements From a Multidisciplinary Expert Panel on the Utilization and Application of a Liver-Specific MRI Contrast Agent (Gadoxetic Acid)

Transcription

1 Gastrointestinal Imaging Review Jhaveri et al. Consensus Statement on the Use of Gadoxetic Acid for Liver MRI Gastrointestinal Imaging Review Kartik Jhaveri 1 Sean Cleary 2 Pascale Audet 3 Fady Balaa 4 Deepak Bhayana 5 Kelly Burak 6 Silvia Chang 7 Elijah Dixon 8 Masoom Haider 9 Michele Molinari 10 Caroline Reinhold 11 Morris Sherman 12 Jhaveri K, Cleary S, et al. Keywords: Eovist, gadoxetate disodium, gadoxetic acid, Gd-EOB-DTPA hepatobiliary, liver, MRI, Primovist DOI: /AJR Received December 17, 2013; accepted after revision June 13, Bayer Inc. provided logistic and communication support for this panel, including the organization of a final panel meeting in Toronto. This article is available for credit. AJR 2015; 204: X/15/ American Roentgen Ray Society Consensus Statements From a Multidisciplinary Expert Panel on the Utilization and Application of a Liver-Specific MRI Contrast Agent (Gadoxetic Acid) OBJECTIVE. This systematic review presents evidence-based consensus statements as reported by a multidisciplinary expert panel (six abdominal radiologists, four hepatobiliary surgeons, and two hepatologists) regarding the use of gadoxetic acid for liver MRI. CONCULSION. Although this review highlights the incremental diagnostic value of hepatobiliary phase imaging with gadoxetic acid enhanced liver MRI in multiple clinical scenarios, there remains a need for further impact studies for some clinical applications, such as hepatocellular carcinoma in cirrhosis. T he constant evolution of technologies in MRI has played an influential role in establishing MRI as an essential modality for the diagnosis and management of liver diseases. Specifically, the introduction of hepatocytespecific contrast agents, such as gadoxetic acid (sold as Primovist [Bayer Healthcare] in Canada, Europe, and Asia and as Eovist [Bayer HealthCare] in the United States), have been shown to improve the detection and characterization of focal liver lesions relative to other imaging modalities, including MRI with conventional extracellular gadolinium based contrast agents [1 4]. Since its initial approval in Europe in 2004 (U.S. approval in 2008 and Canadian approval in 2010), a massive body of literature has been generated from the radiology, hepatology, and surgical specialties investigating the clinical value of gadoxetic acid. As a result, in June 2012, a multidisciplinary expert panel consisting of six abdominal radiologists, four hepatobiliary surgeons, and two hepatologists was conglomerated in Canada, with the goal of providing the medical community with an updated perspective and guidelines regarding gadoxetic acid enhanced liver MRI firmly grounded on the principles of evidence-based clinical practice. The selected panelists were fellowshiptrained subspecialty physicians from tertiary referral centers of care across the country to ensure wide inclusion of national clinical practices and the most directly concerned specialties. The six subspecialty abdominal 1 Division of Abdominal Imaging, Joint Department of Medical Imaging, University Health Network, Mt. Sinai Hospital & Women s College Hospital, University of Toronto, 610 University Ave, Toronto, ON M5G 2M9, Canada. Address correspondence to K. Jhaveri (Kartik.Jhaveri@uhn.ca). 2 Department of Surgery, Division of Hepatobiliary and Pancreatic Surgery, University Health Network, University of Toronto, University Health Network, Toronto, ON, Canada. 3 Department of Radiology, Centre Hospitalier de l Université de Montréal, Montréal, PQ, Canada. 4 Department of Surgery, Division of General Surgery,The Ottawa Hospital, Civic Campus, University of Ottawa, Ottawa, ON, Canada. 5 Department of Radiology, Abdominal Imaging Division, Foothills Hospital, University of Calgary, Calgary, AB, Canada. 6 Departments of Medicine and Oncology, Division of Gastroenterology and Hepatology, University of Calgary, Calgary, AB, Canada. 7 Department of Radiology, Vancouver General Hospital, University of British Columbia, Vancouver, BC, Canada. 8 Division of General Surgery, Hepatobiliary and Pancreatic Surgery, Foothills Medical Centre, University of Calgary,Calgary, AB, Canada. 9 Department of Radiology, Sunnybrook Health Sciences Center, Sunnybrook Research Institute, University of Toronto, Toronto, ON, Canada. 10 Division of General Surgery, Victoria General Hospital, Dalhousie University, Halifax, NS, Canada. 11 Department of Diagnostic Radiology, McGill University Health Center, McGill University, Montréal, QC, Canada. 12 Department of Medicine, Division of Gastroenterology, Toronto General Hospital, University Health Network, University of Toronto, Toronto, ON, Canada. 498 AJR:204, March 2015

2 Consensus Statement on the Use of Gadoxetic Acid for Liver MRI TABLE 1: Classification of Statements According to Levels of Evidence Level of Evidence radiologists were of international or national recognition or had MRI expertise and experience with liver MRI contrast agents for at least 5 years since their fellowship training. Most were in leadership positions in the field of MRI or abdominal imaging or held chairs in radiology with significant clinical MRI expertise or academic merit with critical insight regarding the process. The four surgeons were also fellowship trained in subspecialty hepatobiliary surgical programs and were considered experts in their field regionally or nationally. The two hepatologists on the panel likewise had international or national recognition and essentially represented the eastern and western regions of the country. An extensive comprehensive literature review was performed including the following search terms: Primovist, gadoxetic acid, gadoxetate disodium, gadolinium-ethoxybenzyl-diethylene triamine penta-acetic acid, EOB-DTPA, and MRI. PubMed, MED- LINE, and Cochrane Databases were searched up to August 31, 2013, using English language and human studies as preliminary filters. Reference lists were reviewed manually to identify literature with particular relevance to the predefined subtopics of this review. Subsequent to an exhaustive systematic literature review, which was formulated and provided to the panelists for review, a meeting was held in Toronto, Canada, on September 27, 2013, to discuss and vote on evidencebased statements created by the panel. The statements, originally drafted by the chairs, were first discussed and revised if the need arose, followed by a modified Delphi voting process [5]. In the case of absentees (five in total), comments regarding the originally drafted statements were collected before the meeting and were included in the discussion, and postmeeting voting was executed electronically. Anonymous votes were made on a Likert scale of 1 5 (1, accept completely; 2, accept with some reservation; 3, accept with major reservation; 4, reject with reservation; 5, reject completely) and were collected by a third party experienced in facilitating expert consensus meetings (Mechanisms in Medicine). A predefined acceptance score threshold of 2 (or 80% agreement) was used; all statements qualified after the first voting round and, thus, were included in this article (n = 14). In addition, the statements were also evaluated according to the strength of evidence according to standard practice in evidence-based medicine (Table 1). This review presents the topics, evidence, and respective consensus statements (along with the voting results and mean agreement score) Description Ia Evidence from meta-analysis of randomized controlled trials Ib Evidence from at least one randomized controlled trial Evidence from at least one well-designed controlled trial without randomization IIb Evidence from at least one other type of well-designed quasiexperimental study III Evidence from well-designed nonexperimental descriptive studies, correlation studies, and case studies IV Evidence obtained from expert committee reports or opinions and/or clinical experience of respected authorities) Note Table has been adapted from a U.S. government publication [118] and is in the public domain. in four sections: MRI protocol optimization with gadoxetic acid, liver lesion detection and characterization, biliary applications, and quantitative methods for liver imaging. MRI Protocol Optimization With Gadoxetic Acid Injection Methods and Dosage In patients with normal liver function, approximately 50% of gadoxetic acid is excreted via the hepatobiliary system (the remaining 50% is excreted via the kidneys), which is notably more than other contrast agents exhibiting hepatobiliary excretion, such as gadobenate dimeglumine, which has 0.6 4% hepatobiliary excretion [6]. As a result, and as supported by phase 1 and 2 trials, the recommended dose for gadoxetic acid is mmol/kg of body weight, which is one quarter of the 0.1 mmol/kg of body weight dose recommended for conventional extracellular agents (Table 2). However, the relatively short bolus transit time resulting from TABLE 2: Multidisciplinary Expert Panel Statements 1 5: Protocol Optimization for Hepatobiliary Imaging With Gadoxetic Acid Statement 1: Improved arterial phase enhancement is obtained by MR fluoroscopic or bolus-tracking type triggering technique and either a lower injection flow rate of 1 ml/s or less as opposed to 2 ml/s or higher, or contrast dilution. 2: In patients without cirrhosis or chronic liver disease, hepatobiliary phase acquisition is currently recommended at 20 minutes (after the start of injection) with evolving data suggesting that minutes may be adequate. 3: To improve liver-lesion contrast-to-noise ratio, 3D T1-weighted gradient-echo acquisition for the hepatobiliary phase should be performed with higher flip angles (20 35 ) as opposed to the typical default of : Diffusion-weighted and T2-weighted imaging can be performed after gadoxetic acid administration without compromising diagnostic capability; however, MRCP pulse sequences are currently recommended to be acquired before the contrast agent injection. 5: Insufficient liver enhancement on the hepatobiliary phase images with gadoxetic acid can be seen in patients with moderate liver dysfunction (Child-Pugh class B, model for end-stage liver disease score 11); it may be beneficial to increase dosage to 0.05 mmol/kg of body weight or to acquire delayed hepatobiliary phase images at 30 minutes. In patients with severe liver disease (Child-Pugh class C), the hepatobiliary phase may be of limited utility. Level of Agreement %; %; 3 0%; 4 0%; 5 0% (agreement = 96%) %; 2 8.3%; 3 8.3%; 4 0%; 5 0% (agreement = 95%) 1 50%; 2 50%; 3 0%; 4 0%; 5 0% (agreement = 90%) %; %; 3 0%; 4 0%; 5 0% (agreement = 96%) %; %; 3 0%; 4 0%; 5 0% (agreement = 94%) Level of Evidence AJR:204, March

3 Jhaveri et al. Recommended Optional Dynamic Phases Hepatobiliary Phase Unenhanced 20 s s 2 3 min Contrast-Enhanced min Fig. 1 Recommended workflow and key considerations for gadoxetic acid enhanced liver MRI in patients with normal hepatic function. ADC = apparent diffusion coefficient, GRE = gradient-recalled echo. this reduced dose requires particular attention to the arterial phase acquisition protocol [7, 8]. Modified injection strategies have been proposed to effectively lengthen the arterial bolus transit time, which increases the probability of correctly timing the acquisition to sufficiently fill central k-space, thereby reducing the occurrences of truncation artifacts (e.g., ringing artifacts or Gibbs phenomenon occurring near highly contrasting interfaces seen as bright or dark lines parallel and adjacent to borders of abrupt intensity change). Two injection strategies that have shown favorable results for arterial phase imaging include reducing the injection flow rate and diluting the contrast agent. In a retrospective analysis, Schmid-Tannwald et al. [9] were able to detect significant 2D or 3D T1-weighted in-phase and opposed-phase imaging 2D or 3D T2-weighted fast spin-echo MRCP High-resolution respiratory-triggered 3D sequences should be performed before contrast agent injection. 2D single-shot breath-hold sequences may be acquired immediatedly after portal venous phase imaging. 3D unenhanced T1-weighted fat-saturated GRE Gadoxetic acid injection: 1 2 ml/s, ml saline 3D T1-weighted fat-saturated hepatic arterial phase or triple arterial phase GRE Stretch contrast bolus by using a lower injection rate or dilute with saline. Alternatively, consider triple arterial phase imaging. 3D T1-weighted fat-saturated portal venous phase GRE 3D T1-weighted fat-saturated late venous phase GRE Diffusion-weighted imaging Should be performed after contrast agent injection for time efficiency (there is no change in ADC values). 2D T2-weighted fast spin-echo (with or without fat saturation) Should be performed after contrast agent injection to increase contrast-to-noise ratio and to optimize workflow. 3D T1-weighted fat-saturated hepatocellular phase GRE (20 minutes after injection) Acquire images minutes after contrast agent injection. Consider an increased flip angle. Consider T1-weighted respiratory triggering. 2D T1-weighted imaging (without fat saturation) increments in signal-to-noise ratio and percentage enhancement for the aorta in the arterial phase when using an injection flow rate of 1 ml/s relative to 2 ml/s. These results are further supported by retrospective studies performed by Chung et al. [10] and Haradome et al. [11]. Regarding dilution methods, Motosugi et al. [12] reported an improvement in both overall image quality and lesion-to-liver contrast in the arterial phase images using a diluted method (0.025 mmol/kg of body weight gadoxetic acid diluted with 20 ml of saline, injected at 3 ml/s) relative to a standard injection protocol (0.025 mmol/kg of body weight gadoxetic acid with a 20-mL saline flush, injected at 3 ml/s). Future studies are needed to support the utility of dilution techniques with gadoxetic acid. Finally, although doubling the molar dose of gadoxetic acid to 0.05 mmol/kg of body weight has been proposed and reported without an increase in adverse events, there have been no significant benefits published regarding image quality (portal vein-to-liver contrast) in the population without cirrhosis [13]. With regard to nephrogenic systemic fibrosis, gadoxetic acid, as an ionic linear molecule, has been identified by Health Canada and the U.S. Food and Drug Administration as a relatively low-risk agent. To our knowledge, there have been no confirmed reports of nephrogenic systemic fibrosis linked to the administration of gadoxetic acid; however, we are unable to report any published data about the frequency and outcomes of gadoxetic acid administration in renally impaired patients at high risk for nephrogenic systemic fibrosis. 500 AJR:204, March 2015

4 Consensus Statement on the Use of Gadoxetic Acid for Liver MRI Arterial Phase Timing In addition to injection strategies, the timing of image acquisition after injection is another critical factor for achieving optimal arterial phase image quality, particularly when the protocol includes a single arterial phase. Traditional methods have included fixed delay and test bolus techniques. However, fixed delay methods have been shown to cause unsatisfactory image quality in up to 39% of acquired arterial phase images because of the inability to compensate for interpatient variability (e.g., cardiac output) [14], and test bolus methods can introduce enhancement of the liver parenchyma because of rapid uptake of gadoxetic acid in the hepatocytes, which may negatively affect the visualization of focal liver lesions [15]. Despite the experience and technical requirements, fluoroscopic or bolus-triggering techniques offer more robust timing in that they are inherently patient specific and do not adversely affect the subsequent image quality [16, 17]. When looking at the combined effect of triggering techniques and injection rate to optimize image quality, Haradome et al. [11] found that fluoroscopic triggering with a slow injection rate (1 ml/s) significantly improved arterial phase imaging with gadoxetic acid relative to using a conventional fixed delay method with a fast injection rate (2 ml/s). Hepatobiliary Phase Timing Current recommendations support the acquisition of the hepatobiliary phase at 20 minutes after injection of gadoxetic acid; however, there is a growing body of evidence supporting acquisition as soon as 10 minutes after injection in patients with normal liver function [7, 18 20]. Three recent publications investigating the enhancement dynamics of gadoxetic acid in healthy subjects found that enhancement of the liver parenchyma and common hepatic duct was obvious by 10 minutes after injection of gadoxetic acid [18, 20, 21]. Finally, in a large retrospective study, Motosugi et al. [22] suggested that, when using liver-to-spleen contrast scores as stopping criteria, the hepatobiliary phase could be successfully acquired at 10 minutes in at least 61% of cases. It is important to note that special considerations must be taken into account for patients with cirrhosis and reduced liver function, as described in subsequent sections of this review. Hepatobiliary Phase Flip Angle It has been common practice to simply repeat the hepatobiliary phase with the same acquisition parameters as used in the dynamic phases. However, it has been suggested that acquisition parameters, such as flip angle, should be reexamined for this imaging phase. For instance, by increasing the flip angle in the hepatobiliary phase, the contrast between normal liver parenchyma with gadoxetic acid uptake and excretion (i.e., shorter T1 relaxation times result in greater recovery of longitudinal magnetization) and lesions without such activity (i.e., longer T1 relaxation times result in less recovery of longitudinal magnetization) is effectively increased. Several intraindividual studies have investigated the utility of increasing the flip angle in the hepatobiliary phase using T1-weighted fat-suppressed 3D gradient-recalled echo sequences [23 27]. For instance, significant improvement has been found in the conspicuity and detectability of hypointense lesions in the hepatobiliary phase by increasing the flip angle from to [23 25, 27, 28]. It should be noted that increasing the flip angle comes with the trade-off of increased energy deposition, expressed as specific absorption rate. Because specific absorption rate varies as the square of the flip angle and field strength, caution is recommended when increasing flip angle in the setting of 3-T MRI. In addition to flip angle, it is worth noting that further hepatobiliary phase optimization may be realized by taking advantage of the temporally insensitive nature of this phase (i.e., gadoxetic acid uptake reaches a plateau in the liver parenchyma) relative to the arterial phase (i.e., gadoxetic acid quickly flows through the arterial system). As a result, investigators have noted the advantages of using free-breathing respiratory-triggered T1-weighted acquisitions to improve image resolution and signal-to-noise ratio [29, 30]. Workflow Optimization When using gadoxetic acid, there is a particular opportunity to optimize the image acquisition workflow by leveraging the 10- to 15-minute temporal window between the end of the dynamic phase and the dominance of the hepatobiliary phase (Fig. 1). The acquisition of T2-weighted imaging, diffusionweighted imaging (DWI), and MR cholangiography (MRC) has been investigated in the contrast-enhanced setting. Benndorf et al. [31] reported no significant differences in signal-to-noise ratio, contrastto-noise ratio, or apparent diffusion coefficient values between unenhanced and gadoxetic acid enhanced DWI. These results have been confirmed in other similar studies covering a broad spectrum of patients and hepatic lesions [32 34]. Regarding contrast-enhanced T2-weighted imaging, Choi et al. [33] found no significant differences in qualitative and quantitative image metrics between the unenhanced and gadoxetic acid enhanced T2-weighted images. Similar conclusions were drawn in studies by Saito et al. [32] and Tamada et al. [35], thus confirming the feasibility of performing T2-weighted images after the dynamic imaging and before the hepatobiliary phase. On the other hand, Nakamura et al. [36] found a loss in signal intensity in most biliary structures when performing T2-weighted MRC in the contrast-enhanced setting, thereby suggesting that T2-weighted MRC should not be performed in the period after administering gadoxetic acid. Special Considerations: Parenchymal Enhancement in Cirrhotic Livers The prevailing theory regarding gadoxetic acid hepatocyte uptake is that it is mediated by the organic anion-transporting polypeptide (OATP1 in rats and OATP8 in humans) [37 40]. In the setting of cirrhosis, reduced OATP1 and OATP8 expression has translated into reduced overall uptake of gadoxetic acid in the hepatocytes for rat and human studies, respectively [26, 41 43]. Several reports have shown a significantly higher relative enhancement ratio in the hepatobiliary phase for patients with Child-Pugh class A disease compared with those with Child-Pugh class C disease; however, there are no data to support a difference in enhancement metrics between patients without cirrhosis and patients with Child-Pugh class A and B disease [42, 43]. To compensate for this issue, Motosugi et al. [44] showed an improvement in lesion-toliver contrast ratio in the hepatobiliary phase for patients with Child-Pugh class B disease, when the dose was increased from the standard mmol/kg of body weight to 0.05 mmol/kg of body weight. However, further investigations are warranted to justify the use of an increased dose in patients with cirrhosis. Special Considerations: Protocol Adjustments for Cirrhotic Livers Because of the impaired uptake mechanism of gadoxetic acid in the setting of cirrhosis, the acquisition of hepatobiliary phase beyond the conventional 20 minutes after injection has been suggested [45]. In a prospective study, Tschirch et al. [46] found that AJR:204, March

5 Jhaveri et al. TABLE 3: Multidisciplinary Expert Panel Statements 6 10: Liver Lesion Detection and Characterization hepatobiliary phase image quality was sufficient for healthy control subjects at 20 minutes; however, only 40% of the patients with cirrhosis showed sufficient image quality at both 20 and 30 minutes after injection. The investigators concluded that imaging at 20 minutes after injection with gadoxetic acid is insufficient for patients with liver cirrhosis (model for end-stage liver disease score 11) or those with elevated bilirubin levels ( 1.8 mg/dl or 160 µmol/l). Liver Lesion Detection and Characterization Imaging Modalities in the Management of Liver Metastases Imaging modalities currently used for the management of liver metastases include CT, ultrasound, MRI, and FDG PET (Table 3). A Statement 6a: In the preoperative setting for accurate evaluation of colorectal liver metastases and appropriate surgical planning, gadoxetic acid enhanced liver MRI is recommended because it has superior sensitivity and specificity compared to ultrasound, PET, and CT. 6b: In the assessment of patients with colorectal liver metastases who have been treated with chemotherapy, preoperative imaging with gadoxetic acid may be of particular benefit. 7: For the differentiation of focal nodular hyperplasia from hepatic adenoma, gadoxetic acid enhanced MRI should be considered. 8: Although gadoxetic acid enhanced MRI yields significantly higher diagnostic accuracy and sensitivity compared with multiphase CT for the diagnosis of HCC in cirrhosis, its role in the clinical management of HCC has yet to be defined. 9: There is insufficient evidence supporting cost-effectiveness or outcomes for recommending the utilization of gadoxetic acid enhanced MRI for HCC screening at this time. 10: A significant percentage of nodules with hepatobiliary phase hypoenhancement but atypical enhancement on the dynamic phases have been associated with a diagnosis of HCC or future development of HCC. Biopsy or close surveillance of these lesions is recommended. Note HCC = hepatocellular carcinoma. TABLE 4: Summary of Meta-Analyses Comparing Multiple Modalities for the Detection of Colorectal Cancer Liver Metastases Study (No. of Articles), End Point MRI CT Ultrasound FDG PET Niekel et al. [48] (n = 39) Sensitivity NA 94.1 Specificity NA 95.7 Floriani et al. [47] (n = 25) Sensitivity Specificity Bipat et al. [49] (n = 61) Sensitivity NA 94.6 Specificity NA NA NA NA Note Sensitivities and specificities are given as percentages and are reported on a per-patient basis with all studies using histopathology, intraoperative observation, or follow-up ultrasound as a standard of reference. NA = not applicable. Level of Agreement %; %; 3 0%; 4 8.3%; 5 0% (agreement = 92%) %; %; 3 0%; 4 0%; 5 0% (agreement = 94%) 1 100%; 2 0%; 3 0%; 4 0%; 5 0% (agreement = 100%) %; %; 3 0%; 4 0%; 5 0% (agreement = 96%) 1 100%; 2 0%; 3 0%; 4 0%; 5 0% (agreement = 100%) %; %; 3 0%; 4 0%; 5 0% (agreement = 94%) Level of Evidence number of meta-analyses have been performed summarizing the performance of these modalities for the detection and characterization of liver metastases [47 49] (Table 4). In all cases, MRI proved to be superior to CT; thus, MRI has been suggested to be the preferred first-line modality for evaluating colorectal liver metastases in patients who have not previously undergone therapy and when an extrahepatic evaluation is not desired (e.g., chest) [48]. Despite the relatively high sensitivity and specificity exhibited by FDG PET, the paucity and heterogeneity of the evidence supporting FDG PET have led to its suggested use as a secondline modality. For instance, in the most recent meta-analyses by Niekel et al. [48], it was noted that the data regarding FDG PET (and hybrid FDG PET/CT) were too limited to enable a statistical comparison with the other modalities. Although Floriani et al. [47] reported the advantages in sensitivity for MRI relative to CT when using liver-specific MRI contrast agents in a subgroup of their meta-analysis, there remains a need for quality data comparing the diagnostic performance of gadoxetic acid enhanced MRI relative to conventional extracellular contrast enhanced MRI in this setting. Preoperative Management of Liver Metastases Multimodality imaging plays a significant role in staging colorectal cancer, particularly for the evaluation of liver metastases. The identification of all intrahepatic lesions using highly sensitive imaging is critical to the evaluation of patients with metastatic colorectal cancer, particularly in the assessment of surgical eligibility and appropriate operative planning for potential hepatic resection. Because intrahepatic recurrence is a common cause of postoperative treatment failure, the identification of all potentially malignant lesions and subsequent complete resection are crucial to improving patient outcomes. Relative to dynamic CT, liverspecific MRI contrast agents, in particular gadoxetic acid, have shown superior sensitivity and specificity for the detection and characterization of liver metastases, particularly for small lesions, which makes it a powerful modality in the assessment of patients who are candidates for curative liver resection [1, 4, 50, 51]. Although metastases do not take up gadoxetic acid because of a lack of hepatocyte activity and typically present as hypointense on the hepatobiliary phase, it should be noted that retained uptake of gadoxetic acid has been observed in Ib Ib Ib IV 502 AJR:204, March 2015

6 Consensus Statement on the Use of Gadoxetic Acid for Liver MRI patients with breast cancer, with intratumoral fibrosis resulting in a targetlike appearance on the hepatobiliary phase [52]. In a prospective multicenter intraindividual study, Halavaara et al. [4] were able to show superior characterization of focal liver lesions with gadoxetic acid relative to CT. These findings were later supported by Hammerstingl et al. [1] and Ichikawa et al. [50], who noted the superiority of gadoxetic acid enhanced MRI versus CT for the detection of lesions measuring less than 1 and 2 cm in diameter, respectively. In comparing the diagnostic performance of 64-MDCT and gadoxetic acid enhanced 3-T MRI, Scharitzer et al. [51] concluded that gadoxetic acid enhanced MRI was the preferred choice in the preoperative setting, particularly for the assessment of small lesions. Although gadoxetic acid enhanced MRI is superior to CT, its clinical benefit over MRI with extracellular contrast agents in this setting needs further study, if possible, with intraindividual studies. The diagnostic performance of gadoxetic acid enhanced MRI has been summarized in a recent meta-analysis by Chen et al. [53], who included 13 publications (1900 lesions) evaluating the detection of colorectal cancer liver metastases and reported very high sensitivity and specificity of 93% and 95%, respectively. Imaging of Liver Metastases After Treatment With Chemotherapy Patients who have undergone treatment with systemic chemotherapy may undergo liver parenchyma changes, such as hepatic steatosis or sinusoidal obstruction syndrome, which are known adverse effects of irinotecan- and oxaliplatin-based chemotherapy regimens, respectively. As a result, a fatty (less dense) liver parenchyma will appear less dense on a CT image compared with normal liver parenchyma, thus creating a challenging environment to detect hypovascular liver metastases, such as those from colorectal carcinoma due to a reduction in lesion-to-liver contrast. In a meta-analysis comparing the diagnostic performance of MRI, CT, PET, and hybrid PET/CT in the preoperative setting, Van Kessel et al. [54] found that MRI was the most sensitive technique and should be used as the primary modality to detect metastatic lesions in patients who have undergone chemotherapy. With regard to gadoxetic acid enhanced MRI, Berger-Kulemann et al. [55] performed a prospective analysis in which they concluded that gadoxetic acid enhanced MRI is superior to 64-MDCT in detecting colorectal liver metastases less than 1 cm in diameter during the preoperative staging in patients with liver steatosis. As in the management of metastases for those without previous treatment, a need exists for quality data comparing the diagnostic performance of gadoxetic acid enhanced MRI relative to conventional extracellular contrast enhanced MRI in this posttreatment setting. Focal Nodular Hyperplasia Because of the similar radiologic features and dissimilar natural history and treatment, differentiating between focal nodular hyperplasia (FNH) and hepatocellular adenoma (HCA) is of significant clinical importance. With hepatocyte-specific uptake, it has been shown that gadoxetic acid has the ability to exploit the histologic difference between FNH and HCA via hepatobiliary phase imaging. Despite some occasional distinguishing features, such as the presence of a central scar in FNHs and the presence of high fat and glycogen content in HCAs, these two lesions can share similar features on conventional dynamic imaging. Bieze et al. [56] were able to prospectively show significant advantages in sensitivity in differentiating these lesions when they added the hepatobiliary phase to the dynamic phase using gadoxetic acid enhanced MRI. Gupta et al. [57] compared two hepatocyte-specific agents, gadoxetic acid and gadobenate dimeglumine, and found improved lesion conspicuity in the hepatobiliary phase with gadoxetic acid. With a shorter overall examination time (30 vs minutes), the authors concluded that gadoxetic acid may be a better overall choice for the diagnosis of FNH. Recently, two retrospective analyses reported high accuracy rates of gadoxetic acid enhanced MRI in differentiating FNH and HCA [58, 59]. The combination of hypointensity on hepatobiliary phase images and mild-to-moderate arterial enhancement for HCA versus strong enhancement on arterial phase images and iso- or hyperintensity on hepatobiliary phase images for FNH showed sensitivity and specificity of 83.7% and 100% and 83.8% and 98.5%, respectively [58]. A small percentage of FNHs may not exhibit typical features on hepatobiliary phase imaging attributed to fat content or atypical large central scar, whereas, rarely, HCAs may be isointense to liver on the hepatobiliary phase because of backgroundhepatic steatosis [58]. Hepatic Hemangiomas Hemangiomas appear hypointense relative to the liver parenchyma in the hepatobiliary phase with gadoxetic acid; however, in the dynamic phase images, they may appear different relative to conventional MRI contrast agents because of the rapid parenchymal uptake of gadoxetic acid. With early uptake in the liver parenchyma, the characteristic peripheral nodular enhancement typically observed for hemangiomas in the portal venous phase may be diminished or obscured. Doo et al. [60] recently described the pseudowashout sign in high-flow hemangiomas, which tend to mimic hypervascular tumors. When the pseudowashout pattern is observed, it is recommended that the characteristic bright signal on T2-weighted and arterial phase imaging should be used for the diagnosis of high-flow hemangiomas [60]. In a more innovative approach, Tamada et al. [61] suggested that matching signal intensities between the lesion and portal vein in all phases may be a characteristic pattern for hepatic hemangiomas with gadoxetic acid. Cirrhotic Liver: Hepatocellular Carcinoma Diagnostic imaging of hepatocellular carcinoma The most widely used imaging modalities in hepatocellular carcinoma (HCC) diagnosis are contrast-enhanced MDCT and MRI, which play a significant role in all diagnostic algorithms currently recommended by international guidelines, including the American Association for the Study of Liver Diseases [62, 63], European Association for the Study of the Liver [64], Asian Pacific Association for the Study of the Liver [65], and the Japan Society of Hepatology [66]. Although ultrasound has proven to be relatively insensitive for the diagnosis of HCC in the setting of cirrhosis, it remains as a standard modality for HCC screening, surveillance, and the evaluation of vessel anatomy in the pretransplant setting or for targeted biopsy of a known lesion [67 69]. Despite the fact that MDCT and MRI are generally interchangeable throughout these guidelines, differences in diagnostic performance between gadoxetic acid enhanced MRI and MDCT have been highlighted in multiple reports [70 75]. For instance, Di Martino et al. [73] prospectively compared gadoxetic acid enhanced MRI with multiphasic 64-MDCT and found that gadoxetic acid enhanced MRI AJR:204, March

7 Jhaveri et al. was superior to MDCT in terms of accuracy and sensitivity. Other recent prospective studies comparing gadoxetic acid with MDCT have supported these results, particularly for small HCCs [70, 76, 77]. Within the field of MRI, there have been a number of recent studies investigating the benefits of gadoxetic acid relative to conventional MRI methods. For instance, Bashir et al. [78] recently showed the value of the hepatobiliary phase with gadoxetic acid by showing an improvement in the detection of small (< 1 cm) HCCs. Two recent reports found that gadoxetic acid was significantly more sensitive in detecting HCCs relative to gadopentetate dimeglumine and superparamagnetic iron oxide, respectively [3, 79]. In one of the few studies comparing gadoxetic acid and gadobenate dimeglumine, Park et al. [80] reported a similar diagnostic performance for the two agents, with gadoxetic acid acquired in 20 minutes as compared with 3 hours with gadobenate dimeglumine. Hepatocellular carcinoma screening In the current international guidelines, ultrasound is the most widely endorsed modality for screening and surveillance of patients at risk for HCC; however, suboptimal performance characteristics of ultrasound in combination with its high dependency on operator proficiency have led to an interest in exploring the utility of CT and MRI in the surveillance setting. As a result, the Japan Society of Hepatology guidelines have included dynamic CT, dynamic MRI, and gadoxetic acid enhanced MRI as surveillance tools for patients classified as super-high risk (i.e., those with hepatitis B or C related cirrhosis) [66]. At this point, the explicit integration of gadoxetic acid enhanced MRI as a tool for HCC surveillance is not supported because of the lack of cost-benefit analysis with this modality. Staging of Hepatocellular Carcinoma The most successful therapies for HCC designed with curative intent are surgical resection, transplantation, and ablation [81, 82]. The detection all HCC lesions during the staging process is key to avoiding recurrence after liver resection and, to a lesser extent, transplantation. Bashir et al. [78] found that lesion detection was substantially higher for small HCCs when they included the hepatobiliary phase with gadoxetic acid enhanced MRI, as opposed to conventional dynamic MRI. These findings are consistent with those of a study by Baek et al. [83], who were able to show a statistically significant improvement in detecting HCC lesions measuring less than 1 cm in diameter for gadoxetic acid enhanced MRI relative to MDCT. At this point, there remains a need for evidence from a well-designed multicenter prospective study assessing the utility of gadoxetic acid enhanced MRI in staging of HCC in North American patients with cirrhosis. Moreover, given that the current North American guidelines for the diagnosis and management of HCC are based on extracellular gadolinium based MRI contrast agents and CT, particularly with regard to liver transplantation [84], the precise role of gadoxetic acid in this scenario is currently uncertain. In addition to the detection of HCC, the histopathologic grade of these lesions can play a significant role in deciding treatment strategy; thus, many efforts have focused on estimating the HCC histopathologic grade using gadoxetic acid [85 88]. For instance, An et al. [86] reported that an increase in arterial enhancement with gadoxetic acid in combination with restricted diffusion on DWI is suggestive of higher-grade HCC. Moreover, Choi et al. [85] reported that a low contrast enhancement ratio on gadoxetic acid enhanced hepatobiliary phase images may have the potential to be used as a marker of poor prognosis for HCC. Despite encouraging results, further investigation is warranted on this topic before supporting reliance on liver MRI with gadoxetic acid for the purpose of HCC grading. Cirrhotic Nodules The development of HCC typically involves a multistep progression from regenerative nodules to low-grade dysplastic nodules to high-grade dysplastic nodules and, finally, to early HCC and overt HCC. Differentiating between premalignant nodules and early HCC is critical to patient management; however, challenges may arise when using conventional dynamic imaging with CT or MRI because early HCCs may resemble benign dysplastic nodules present in cirrhotic livers. Rhee et al. [89] described the following diagnostic criteria for discriminating early HCCs from benign dysplastic nodules: a size greater than or equal to 1.5 cm diameter, hypointensity on T1-weighted imaging, hyperintensity on T2-weighted imaging and DWI, hyperenhancement on the arterial phase, washout on the portal venous, and hypointensity on the hepatobiliary phase. Chou et al [90]. reported that the hepatobiliary phase with gadoxetic acid may provide incremental value to conventional dynamic MRI for the characterization of focal lesions in the cirrhotic liver. As a result of evolving data, the Japan Society of Hepatology guidelines have explicitly recommended the use of gadoxetic acid enhanced MRI for the differentiation of HCCs and dysplastic nodules [66]. Although results have shown the potential value of gadoxetic acid in this setting, it is important to note that challenges may still arise because of benign nodules presenting as hypointense or because of early HCCs presenting as isoor hyperintense. There is an ongoing debate and lack of uniform acceptance among the radiologic community regarding the use of gadoxetic acid in the setting of cirrhosis related to potentially reduced arterial phase enhancement, as well as challenges from assessing venous or intermediate phase washout compared with extracellular gadolinium based contrast agents. Also, the current diagnostic criteria for definite HCC for liver transplantation eligibility and the Liver Imaging Reporting and Data System in North America are based on the utilization of extracellular contrast agents, not gadoxetic acid. Thus, there is insufficient evidence at this time to recommend its use for this indication. Impact on Health Economics In a European health economics analysis by Zech et al. [91], gadoxetic acid enhanced MRI was associated with significant cost savings relative to extracellular gadolinium enhanced MRI and MDCT by improving preoperative planning and reducing intraoperative changes. Although these data provide some justification for the incremental cost of gadoxetic acid enhanced liver MRI, given the complexity of cost estimations in health care economics, caution is advised when considering these data in other health care systems, and there may be a need to validate the results in other countries. Biliary Applications Imaging Biliary Anatomy in the Setting of Liver Donors The frequency of vascular or biliary complications in healthy donors as a result of partial hepatectomy have been reported in the range of 2 32% [92] (Table 5). In a recent study by Mangold et al. [93] evaluating living liver donors, it was found that the small second- and third-order intrahepatic bile ducts were more accurately depicted when using gadoxetic acid enhanced images relative to conventional T2-weighted MRC sequences. 504 AJR:204, March 2015

8 Consensus Statement on the Use of Gadoxetic Acid for Liver MRI TABLE 5: Multidisciplinary Expert Panel Statements 11 and 12: Biliary Applications The well-recognized complementary nature of conventional T2-weighted MRC and contrast-enhanced T1-weighted MRC has led to a number of studies investigating the value of combining these two techniques for the visualization of the biliary system [94 97]. Carlos et al. [97] were able to show significant improvements when using both techniques with gadoxetic acid relative to each modality alone. Moreover, those same authors concluded that a 20-minute delay after gadoxetic acid injection was sufficient delay for adequate biliary enhancement [98]. Bile Leaks In addition to visualizing biliary anatomy, gadoxetic acid enhanced MRC has shown value in the detection of postoperative bile leaks, including complications after cholecystectomy. In the setting of bile leaks, a positive diagnosis may be considered in gadoxetic acid enhanced hepatobiliary phase images via homogeneous enhancement of both the biliary tree and areas where enhancement is visible outside the bile ducts. In the most recent study by Kantarci et al. [96], the sensitivity, specificity, and accuracy of locating biliary leaks were significantly higher when using T2-weighted MRC and gadoxetic acid enhanced MRC together, compared with either one alone. Castellanos et al. [99] found that gadoxetic acid was able to detect all bile leaks (100%) that appeared 20 minutes after administration of the contrast agent and thus was a highly reliable technique and advantageous over invasive strategies. It should be noted that the aforementioned use of gadoxetic acid in assessing biliary anatomy and bile leaks is under the umbrella Statement 11: The addition of gadoxetic acid enhanced MRC to conventional T2-weighted MRC increases the accuracy in displaying biliary anatomy, particularly in the assessment of potential living-related liver donors and in the delineation of postoperative bile leaks. 12: Gadoxetic acid enhanced MRI for cholangiocarcinoma may offer improved tumor margin delineation, characterization of associated hepatic lesions, and provide accurate relationship of the tumor to bile ducts for surgical resection planning. Note MRC = MR cholangiography. of off-label use given the regulatory approval of gadoxetic acid for the detection and characterization of focal liver lesions. Biliary Tumors To date, resection remains the treatment of choice for cholangiocarcinoma, with ultrasound, MDCT, MRC, FDG PET, or ERCP being the most common modalities used to help determine resectability by defining the proximal and distal extent of biliary involvement of the tumor and the detection of vascular involvement. Péporté et al. [100] found that all cholangiocarcinomas presented as hypointense in the late and hepatobiliary phases relative to the liver parenchyma, with the most conspicuous presentation in the hepatobiliary phase. They also reported that both contrast-to-noise ratio and signal-tonoise ratio of these lesions were significantly higher for the gadoxetic acid enhanced images relative to unenhanced images [100]. With improved conspicuity of cholangiocarcinoma lesions, this study was able to show the incremental value of the hepatobiliary phase provided by gadoxetic acid enhanced MRI. Finally, a recent retrospective analysis concluded that gadoxetic acid enhanced MRC is a reliable method for assessing bile duct tumor resectability, with an accuracy up to 83.5% [101]. Quantitative Methods for Liver Imaging One area of research that has emerged in recent years is the quantitative analysis of hepatic function using gadoxetic acid enhanced MRI (Table 6). This is achieved by exploiting gadoxetic acid s uptake mechanism in the liver cells, which is governed Level of Agreement %; %; 3 0%; 4 8.3%; 5 0% (agreement = 92%) %; %; 3 8.3%; 4 0%; 5 0% (agreement = 85%) Level of Evidence by the number and functional capacity of the hepatocytes themselves. The interest in gadoxetic acid enhanced MRI is based on the potential to quantify hepatic function on a segmental level, which could serve as a highly valuable tool in providing an accurate preoperative prediction of postoperative liver function and regeneration capacity [102, 103]. A wide variety of techniques have been proposed in this evolving area of MRI, ranging from biomarker correlation to deconvolution models and T1 signal mapping techniques [ ]. In addition, recent advances in functional MRI have led to a growing interest in the assessment of hepatic fibrosis using advanced methods, such as DWI [111], MR elastography [112], MR spectroscopy [113], and contrast-enhanced MRI with hepatocyte-specific contrast agents, such as gadoxetic acid [ ]. With regard to gadoxetic acid enhanced MRI, a common approach to staging hepatic fibrosis has been based on signal-intensity metrics in the hepatobiliary phase. Preliminary investigations have shown good correlation between traditional biochemical measures of synthetic hepatic function (model for end-stage liver disease and Childs-Pugh score) and several measures of gadoxetic acid kinetics, including hepatic extraction fraction, median transit time, and input-relative blood flow [107]. These techniques appear promising but are currently in the research realm awaiting translation into clinical practice. Limitations We would like to acknowledge some limitations to this article. Our work was not a meta-analysis of the existing literature but TABLE 6: Multidisciplinary Expert Panel Statement 13: Quantitative Methods for Liver Function Estimation 1b Statement 13: Gadoxetic acid enhanced liver MRI is an evolving technique with potential for noninvasive quantification of liver function and staging of hepatic fibrosis. Level of Agreement %; %; 3 0%; 4 0%; 5 0% (agreement = 94%) Level of Evidence AJR:204, March

9 Jhaveri et al. rather a systematic review firmly grounded on evidence in the literature, and the consensus statements were arrived at on the basis of evidence, not expert opinions. Perhaps in the need to keep our review evidence based, some of our consensus statements may read as rather conservative and not progressive enough for some readers, particularly the section on HCC. However, our primary aim was to provide indications and statements that are supported by a strong body of literature, not evolving applications. We have, however, briefly discussed evolving applications here as well. Conclusion Existing and rapidly evolving evidence in the field of gadoxetic acid enhanced MRI provides perspective regarding its established, potential, and future applications. It is clear that the additional hepatobiliary phase gained when administering gadoxetic acid offers incremental value in a number of indications and clinical settings; however, it is important for users to actively optimize their imaging protocols to fully realize this value. This article has provided an upto-date balanced evidence-based review, including 14 consensus statements from a multidisciplinary panel of physicians, that was designed to provide guidance to the user on how and when to use gadoxetic acid in the setting of hepatobiliary imaging. Acknowledgment We thank Tyler Hughes for his assistance in the preparation of the manuscript. References 1. Hammerstingl R, Huppertz A, Breuer J, et al. Diagnostic efficacy of gadoxetic acid (Primovist)- enhanced MRI and spiral CT for a therapeutic strategy: comparison with intraoperative and histopathologic findings in focal liver lesions. Eur Radiol 2008; 18: Filippone A, Blakeborough A, Breuer J, et al. Enhancement of liver parenchyma after injection of hepatocyte-specific MRI contrast media: a comparison of gadoxetic acid and gadobenate dimeglumine. J Magn Reson Imaging 2010; 31: Park G, Kim YK, Kim CS, Yu HC, Hwang SB. Diagnostic efficacy of gadoxetic acid-enhanced MRI in the detection of hepatocellular carcinomas: comparison with gadopentetate dimeglumine. Br J Radiol 2010; 83: Halavaara J, Breuer J, Ayuso C, et al. Liver tumor characterization: comparison between liver-specific gadoxetic acid disodium-enhanced MRI and biphasic CT a multicenter trial. J Comput Assist Tomogr 2006; 30: Williams PL, Webb C. The Delphi technique: a methodological discussion. J Adv Nurs 1994; 19: Bracco Diagnostics Canada. MultiHance (gadobenate dimeglumine injection) product monograph. Bracco Diagnostics website. imaging. bracco.com/ us-en/products-and-solutions/contrast-media/ multihance. Published Accessed October 31, Hamm B, Staks T, Mühler A, et al. Phase I clinical evaluation of gadoxetic acid as a hepatobiliary MR contrast agent: safety, pharmacokinetics, and MR imaging. Radiology 1995; 195: Reimer P, Rummeny E, Shamsi K, et al. Phase II clinical evaluation of gadoxetic acid: dose, safety aspects, and pulse sequence. Radiology 1996; 199: Schmid-Tannwald C, Herrmann K, Oto A, Panteleon A, Reiser M, Zech C. Optimization of the dynamic, gadoxetic acid-enhanced MRI of the liver: the effect of the injection rate. Acta Radiol 2012; 53: Chung SH, Kim MJ, Choi JY, Hong HS. Comparison of two different injection rates of gadoxetic acid for arterial phase MRI of the liver. J Magn Reson Imaging 2010; 31: Haradome H, Grazioli L, Tsunoo M, et al. Can MR fluoroscopic triggering technique and slow rate injection provide appropriate arterial phase images with reducing artifacts on gadoxetic acid-dtpa (gadoxetic acid)-enhanced hepatic MR imaging? J Magn Reson Imaging 2010; 32: Motosugi U, Ichikawa T, Sou H, et al. Dilution method of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (gadoxetic acid)-enhanced magnetic resonance imaging (MRI). J Magn Reson Imaging 2009; 30: Feuerlein S, Boll DT, Gupta RT, Ringe KI, Marin D, Merkle EM. Gadoxetate disodium enhanced hepatic MRI: dose-dependent contrast dynamics of hepatic parenchyma and portal vein. AJR 2011; 196:[web]W18 W Ringe KI, Husarik DB, Sirlin CB, Merkle EM. Gadoxetate disodium enhanced MRI of the liver. Part 1. Protocol optimization and lesion appearance in the noncirrhotic liver. AJR 2010; 195: Haradome H, Grazioli L, Tsunoo M, et al. Can MR fluoroscopic triggering technique and slow rate injection provide appropriate arterial phase images with reducing artifacts on gadoxetic acid-dtpa (gadoxetic acid)-enhanced hepatic MR imaging? J Magn Reson Imaging 2010; 32: Wilman AH, Riederer SJ, King BF, Debbins JP, Rossman PJ, Ehman RL. Fluoroscopically triggered contrast-enhanced three-dimensional MR angiography with elliptical centric view order: application to the renal arteries. Radiology 1997; 205: Huston J, Fain SB, Riederer SJ, Wilman AH, Bernstein MA, Busse RF. Carotid arteries: maximizing arterial to venous contrast in fluoroscopically triggered contrast-enhanced MR angiography with elliptic centric view ordering. Radiology 1999; 211: Dahlström N, Persson A, Albiin N, Smedby O, Brismar TB. Contrast-enhanced magnetic resonance cholangiography with Gd-BOPTA and gadoxetic acid in healthy subjects. Acta Radiol 2007; 48: Motosugi U, Ichikawa T, Sou H, et al. Liver parenchymal enhancement of hepatocyte-phase images in gadoxetic acid-enhanced MR imaging: which biological markers of the liver function affect the enhancement? J Magn Reson Imaging 2009; 30: Van Kessel CS, Veldhuis WB, van den Bosch MA, van Leeuwen MS. MR liver imaging with Gd-EOB-DTPA: a delay time of 10 minutes is sufficient for lesion characterisation. Eur Radiol 2012; 22: Brismar TB, Dahlstrom N, Edsborg N, Persson A, Smedby O, Albiin N. Liver vessel enhancement by Gd-BOPTA and Gd-EOB-DTPA: a comparison in healthy volunteers. Acta Radiol 2009; 50: Motosugi U, Ichikawa T, Tominaga L, et al. Delay before the hepatocyte phase of Gd-EOB-DT- PA-enhanced MR imaging: is it possible to shorten the examination time? Eur Radiol 2009; 19: Bashir MR, Husarik DB, Ziemlewicz TJ, Gupta RT, Boll DT, Merkle EM. Liver MRI in the hepatocyte phase with gadolinium-eob-dtpa: does increasing the flip angle improve conspicuity and detection rate of hypointense lesions? J Magn Reson Imaging 2012; 35: Bashir MR, Merkle EM. Improved liver lesion conspicuity by increasing the flip angle during hepatocyte phase MR imaging. Eur Radiol 2011; 21: Haradome H, Grazioli L, Al manea K, et al. Gadoxetic acid disodium-enhanced hepatocyte phase MRI: can increasing the flip angle improve focal liver lesion detection? J Magn Reson Imaging 2012; 35: Tamada T, Ito K, Higaki A, et al. Gd-EOB-DT- PA-enhanced MR imaging: evaluation of hepatic enhancement effects in normal and cirrhotic livers. Eur J Radiol 2011; 80:e311 e Tamada T, Ito K, Yamamoto A, et al. Hypointense hepatocellular nodules on hepatobiliary phase of Gd-EOB-DTPA-enhanced MRI: can increasing the flip angle improve conspicuity of lesions? J 506 AJR:204, March 2015