Review of hepatobiliary contrast agents: Current applications and challenges

Watch a video presentation of this article

Abbreviations

EMA

European Medicines Agency

FDA

US Food and Drug Administration

FNH

focal nodular hyperplasia

GBCA

gadolinium‐based contrast agent

HA

hepatic adenoma

hbGBCA

hepatobiliary GBCA

MRC

magnetic resonance cholangiography

MRCP

magnetic resonance cholangiopancreatography

MRI

magnetic resonance imaging

NSF

nephrogenic systemic fibrosis

The group of hepatobiliary contrast agents comprises two gadolinium‐based contrast agents (GBCAs) that show a specific amount of hepatocellular uptake and biliary excretion. Gadobenate dimeglumine (Gd‐BOPTA; Bracco Diagnostics) and gadoxetic acid (Gd‐EOB‐DTPA; Bayer Healthcare) are US Food and Drug Administration (FDA)‐approved agents for magnetic resonance imaging (MRI) and for detection and characterization of liver lesions, respectively. A third agent with hepatobiliary properties, gadofosveset trisodium, has recently been discontinued.

The success of these so‐called hepatobiliary contrast agents (hbGBCAs) is essentially related to satisfying the fundamental motive in MRI—that is, the pursuit to structurally characterize tissues or lesions. Specifically, MRI's high intrinsic soft tissue contrast with its ability to enhance specific tissue characteristics is further amended by the ability of hbGBCAs to detect hepatocytes and biliary canaliculi. In combination, the use of hbGBCAs in MRI is a powerful diagnostic instrument.

hbGBCAs carry the dual properties of acting like “conventional” GBCAs during the first pass and early phases of contrast enhancement, and by revealing the characteristic delayed or hepatobiliary enhancement. In the initial phases after injection, their contrast enhancement is dominated by the distribution in vessels and the extracellular space like in conventional GBCAs. However, unlike other GBCAs, hbGBCAs are then taken up by functional hepatocytes. As a result, there is a persistent enhancement in hepatocytes that allows for identifying lesions of hepatocellular origin. In addition, the biliary excretion allows for both identifying lesions that contain biliary canaliculi and enhancing the biliary excretion system for magnetic resonance cholangiography (MRC). Of note, there is a relative difference in hepatobiliary excretion between both agents (gadoxetic acid: ∼50%, gadobenate dimeglumine: 4‐5%) leading to biliary enhancement starting at 15 to 25 minutes after injection (gadoxetate acid) and 60 to 90 minutes (gadobenate dimeglumine), which has an impact on workflow considerations. Also, there are reports that add to the author's experience that there are qualitative differences in imaging properties during the early phases of contrast enhancement that are dominated by the extracellular properties of both agents related to the earlier extraction of gadoxetic acid from the extracellular space. However, the published data are not yet conclusive. Some authors have therefore suggested to increase the dose of gadoxetic acid.1, 2

Although the advantage of hbGBCAs over other GBCAs in characterizing hepatocellular carcinoma and differentiating HCC from dysplastic regenerative nodules is a matter of ongoing research, their strength can truly be appreciated in the differential diagnosis of hepatic adenomas (HAs) and focal nodular hyperplasia (FNH).3 Although both are benign tumors of hepatocellular origin, HAs show a variable appearance, may show hemorrhage, and have been associated with the risk for malignant transformation. The key for differentiating both lesions is the presence of biliary structures: Although FNHs contain often malformed biliary canaliculi, they are typically absent in HAs. As shown in Figs. 1 and 2, this leads to comparable imaging features during the early phases of contrast enhancement, but distinct differences in the delayed or hepatobiliary phase. In the delayed phase, the FNH retains contrast enhancement and will therefore reveal an isointense to hyperintense signal in comparison with the liver parenchyma. HAs, however, lack biliary structures and will typically be hypointense during the delayed phase as compared with healthy liver parenchyma. Similarly, lesions without hepatocytes such as metastases can readily be differentiated because of the lack of hepatocellular uptake. An in‐depth overview can be found in Frydrychowicz et al.2 Nevertheless, care should be taken not to blindly trust hbGBCA‐enhanced magnetic resonance images only. The entire spectrum of MRI sequences including diffusion‐weight imaging should be used to narrow in on the most likely diagnosis. Especially in the presence of hemorrhage or inflammation, both FNH and HA can present with very variable MRI characteristics, and biopsy may be required.

Figure 1.

MRI of the liver using the hbGBCA gadoxetic acid in a 23‐year‐old woman sent to characterize an incidental liver mass. A mass with a slight edema (A) in the right liver lobe can be appreciated. Typical hbGBCA enhancement features of this histologically proven FNH (Mohajer type 1 pattern3) can be recognized: hypervascularity on the arterial‐phase image with a central scar (B), decreasing enhancement that is isointense to hyperintense to the liver in the portal venous and venous phases (C and D), and persistent enhancement during the delayed or hepatobiliary phase [E and F, respectively; shown are scans at 20 (E) and 30 minutes (F) after injection].

Figure 2.

MRI characteristics of a histologically proven HA in a 31‐year‐old woman with gadoxetic acid–enhanced MRI. Although this lesion shows similar edema (A) and intense arterial‐phase enhancement (B) as compared with the lesion in Fig. 1, it is isointense to slightly hypointense during the portal venous (C) and venous phases (D), and does not exhibit contrast enhancement during the delayed or hepatobiliary phase (E and F). The asterisk (*) marks an unexplained, transient arterial enhancement of Couinaud liver segment 5.

There are sparse data regarding the clinical value of hbGBCA‐based MRC and the comparison with three‐dimensional T₂‐weighted magnetic resonance cholangiopancreatography (MRCP). The available data suggest that both MRCP and MRC should be used complementary.4 Especially in postoperative courses, the biliary excretion of hbGBCAs can be advantageous in questions such as biliary leakage (see Fig. 3) or patency of biliodigestive anastomosis. In the latter, contrast enhancement in the duodenum after hbGBCA application is able to confirm patency of the anastomosis, functional information that is not available by MRCP alone. Notably, when performing hbGBCA‐MRC, adaptation of the flip angle to the applied contrast agent and field strength has shown to be beneficial to the resulting image quality.1, 5

Figure 3.

Detection of biliary leakage in a 35‐year‐old male patient after surgery because of traumatic liver laceration. The patient returned several days after having been discharged free of symptoms with unspecific abdominal symptoms. An ultrasound survey picked up free fluid next to the liver. The patient had MRI, which confirmed prehepatic fluid (A, asterisks) adjacent to the falciform ligament where the rupture had been treated (white arrow). During delayed‐phase imaging (B, 25 minutes; C, 45 minutes after injection of gadoxetic acid) biliary leakage was confirmed by extrahepatic pooling of contrast agent (+). Note the beginning of biliary contrast enhancement at 25 minutes after injection (open arrows) and a dilated bile duct in Couinaud liver segment 2 (arrowhead).

The ability to exploit the functional information using hbGBCAs has spurred state‐of‐the‐art research. Under the premise that functional hepatocytes are necessary for hepatocellular uptake of these GBCAs, liver enhancement is correlated with hepatocytic function. Katsube et al.6 have shown Look‐Locker‐type T₁‐mapping to evaluate T₁‐shortening as the effect of liver enhancement at different time points after gadoxetic acid injection. They were able to quantitatively differentiate patients with normal liver function from those with various stages of liver cirrhosis. With the increased availability of mapping sequences, this approach has recently been revisited by various authors and is promising for future comprehensive liver imaging protocols.

Challenges potentially relevant to hbGBCA MRI are associated with the clearance of each agent from the body. To one end, nephrogenic systemic fibrosis (NSF) has led to reconsideration of any indiscriminate application and dosing of GBCAs. This topic has fortunately calmed down significantly using caution in selecting patients with sufficient renal clearance. Also, the preference of so‐called macrocyclic instead of linear GBCA compounds, as well as attention towards the molecular charge, may have contributed to decreased NSF incidence.7 Although there is an ongoing debate on this complex matter, it has to be noted that even though both hbGBCAs are linear compounds, they have been associated with few, if any, NSF incidents.8 In light of reports on compensatory mechanisms regarding the excretion pathway being upscaled in restricted renal function,9 the hepatobiliary excretion may be a decisive factor.

To the other end, the debate on linear versus macrocyclic GBCAs has just recently been revisited following reports on the deposition of gadolinium in the dentate nucleus.10 Whereas in Europe, the European Medicines Agency (EMA) has recommended to suspend linear GBCAs to the European Commission despite no adverse health effects having been identified,11 the FDA has stated “to continue assessing all GBCAs' safety and to reevaluate GBCAs pending further testing.”12 Interestingly, the EMA's recommendations to suspend linear GBCAs excluded gadoxetic acid and suggested to restrict the use of gadobenate dimeglumine to liver imaging, most likely taking into account the hbBGCA's unmatched diagnostic properties and excretion pathways.

Another noteworthy challenge when using gadoxetic acid is an unphysiological mechanism not yet fully understood: There are reports on shortened breathhold durations in some patients due to a transient dyspnea linked to the injection of gadoxetic acid.13 This phenomenon has since been confirmed by other authors and can explain the deterioration of arterial‐phase breathhold imaging in some cases and should be taken into account when setting up dynamic‐phase imaging protocols when using gadoxetic acid. Consequently, this may have unfavorable effects in arterial‐phase imaging and especially for HCC detection and characterization, and may have influenced, at least to some extent, the published results on HCC detection rates. Also, variable signal characteristics of some HCCs in the hepatobiliary phase should be taken into account. Typically, HCCs present hypointense on T₁‐weighted fat‐saturated images. In some cases, delayed‐phase T₁‐weighted hyperintensity can be observed that has been attributed to different, increased expression rates of the transporters necessary for the hepatobiliary excretion pathway.14

In summary, hepatobiliary gadolinium‐based contrast agents are powerful diagnostic instruments that satisfy routine diagnostic imaging demands and can act as problem‐solvers in situations such as liver lesion differentiation or biliary leakage. Their unique diagnostic potential may be further augmented by T₁ mapping of liver function. Also, the proposed compensatory excretion mechanism is potentially beneficial to patients with decreased renal function.