JOURNAL OF MAGNETIC RESONANCE IMAGING 25:775–782 (2007)
Original Research
Focal Nodular Hyperplasia: Intraindividual Comparison of Dynamic Gadobenate Dimeglumineand Ferucarbotran-Enhanced Magnetic Resonance Imaging Daniele Marin, MD,1 Riccardo Iannaccone, MD,1* Andrea Laghi, MD,1 Carlo Catalano, MD,1 Takamichi Murakami, MD, PhD,2 Masatoshi Hori, MD, PhD,2 Tonsok Kim, MD,2 and Roberto Passariello, MD1 Purpose: To intraindividually compare the enhancement pattern of focal nodular hyperplasia (FNH) after dynamic administration of two bolus-injectable liver-specific MR contrast agents, ferucarbotran and gadobenate dimeglumine.
Key Words: liver, MR; liver, contrast agents; benign liver neoplasm; focal nodular hyperplasia; comparative study J. Magn. Reson. Imaging 2007;25:775–782. © 2007 Wiley-Liss, Inc.
Materials and Methods: A total of 19 patients with 24 FNHs underwent gadobenate dimeglumine- and ferucarbotran-enhanced MRI during the hepatic arterial-dominant phase (HAP; 25 seconds), the portal-venous phase (PVP; 60 seconds), and the equilibrium phase (EP; 180 seconds). Hepatospecific phases were acquired on T1-weighted images 120 minutes after gadobenate dimeglumine administration, and on T2-weighted images 10 minutes after ferucarbotran administration. Lesion enhancement was independently analyzed by two observers. The kappa statistic was determined to evaluate the agreement between the enhancement patterns of the lesions. Results: On gadobenate dimeglumine– enhanced MR images during HAP, PVP, and EP, FNHs were: hyperintense (24/20/13); isointense (0/4/11); and hypointense (0/0/0). On ferucarbotran-enhanced MR images during HAP, PVP, and EP, FNHs were: hyperintense (2/0/0); isointense (16/ 9/14); and hypointense (6/15/10). Overall, poor agreement between both contrast agents was observed. During the hepatospecific phases, most (20/24; 83%) FNHs showed a typical enhancement pattern during the delayed hepatospecific phase. Conclusion: The dynamic enhancement pattern of FNHs is significantly different between gadobenate dimeglumine– and ferucarbotran-enhanced MRI. With respect to hepatospecific phase, the majority of FNHs showed a typical behavior on both contrast agents.
1 Department of Radiological Sciences, University of Rome “La Sapienza”, Rome, Italy. 2 Department of Radiology, Osaka University Graduate School of Medicine, Osaka, Japan. *Address reprint requests to: R.I., Department of Radiological Sciences, University of Rome “La Sapienza”, Viale Regina Elena 324, Rome 00161, Italy. E-mail:
[email protected] DOI 10.1002/jmri.20885 Published online 8 March 2007 in Wiley InterScience (www.interscience. wiley.com).
© 2007 Wiley-Liss, Inc.
FOCAL NODULAR HYPERPLASIA (FNH) is a benign tumor of the liver that generally occurs in childbearing and middle-aged asymptomatic women. Rather than a neoplastic process, it is considered to be a hyperplastic response to a congenital or acquired anomaly of the arterial blood supply leading to focal hyperperfusion of the hepatic parenchyma (1). Due to the lack of malignant potential and the extremely low complication-rate, such as rupture and hemorrhage (2), FNH usually warrants conservative management (3,4). With current advances in modern cross-sectional imaging, an increasing number of FNHs are incidentally detected. Multiphasic helical computed tomography (CT) and MRI generally allow confident diagnosis of typical FNH lesions (5,6). Bright and homogeneous lesion enhancement on the hepatic arterial phase (HAP) and, to a lesser extent, the presence of a central fibrous scar, are the most reliable CT and MR signs for FNH detection and characterization (5,6). However, atypical imaging features (6 –9), multifocal distribution, and association with other hypervascular liver lesions, such as hemangioma (10) and adenoma (11), occasionally make the diagnosis of FNH a clinicoradiological challenge. By providing simultaneous morphologic and functional information, hepatospecific MR contrast agents could lead to improvement in the detection and characterization of focal hepatic lesions. Gadobenate dimeglumine (MultiHance威; Bracco, Milan, Italy) is a liver-specific paramagnetic gadoliniumbased MR contrast agent with a vascular-interstitial distribution in the first minutes after bolus injection, followed by delayed hepatobiliary excretion (12). Some recent studies have emphasized the value of both dynamic and delayed gadobenate dimeglumine T1-weighted
775
776
images for the identification and characterization of FNH (13–15). Ferucarbotran (Resovist威; Schering AG, Berlin, Germany) is a clinically safe, fast bolus-injectable, reticuloendothelial system (RES)-targeted MR contrast agent, composed of carboxydextran-coated superparamagnetic iron oxide (SPIO) particles, available for liver imaging in most European countries as well as some countries in Asia. Due to the high r1 and r2 relaxivities in blood, this contrast agent simultaneously enables, during a single examination, the performance of a T1weighted dynamic perfusion study with timing strategies comparable to those used with commercially available paramagnetic gadolinium chelates, followed by T2and T2*-weighted delayed RES-specific accumulation phase imaging (16,17). To our knowledge, only a few feasibility studies have systematically evaluated the T1 effects of bolus-injected ferucarbotran during dynamic MRI (16 –22). Specifically, at present, there are no comprehensive data about the dynamic perfusion pattern of FNH on ferucarbotran-enhanced T1-weighted images. The purpose of this prospective study was to intraindividually compare the enhancement pattern of FNH after dynamic administration of two different bolusinjectable, liver-specific MR contrast agents, ferucarbotran and gadobenate dimeglumine.
MATERIALS AND METHODS Study Population The study received approval from our Institutional Review Board and followed the Declaration of Helsinki principles (23). We performed a Boolean search for the words “focal,” “nodular,” and “hyperplasia” in the medical records, radiology files, and pathology databases at our Institution between January 1, 2000 and January 30, 2004, and identified 120 consecutive patients. We selected 29 cases in which the diagnosis was specifically established by means of pathological confirmation or based on strict clinicoradiological criteria (see below). Of these patients, 19 (16 women, three men; mean age ⫽ 40 years; age range ⫽ 10 –59 years) with a total of 24 FNH lesions accepted enrollment in the study and formed our final study population. A total of 16 patients had one lesion, two patients had two lesions, and one patient had four lesions. Nine women had a previous history of oral contraceptive use for an average of 4.5 ⫾ 2 years; none of the men were using steroids. None of these patients had a history of chronic liver disease or viral hepatitis infection. Exclusion criteria included: a referred previous history of hypersensitivity to any metals and/or chelates of gadolinium, clinical evidence of congestive heart failure (New York Heart Association class III or IV), a body weight of more than 145 kg, iron overload or ongoing supplementation therapy, pregnancy or lactation, and any contraindication to MRI (i.e., pacemaker, claustrophobia, electronic devices, or ferromagnetic vascular clips). Patients were prospectively recruited between January 1, 2004 and January 15, 2005 and written informed
Marin et al.
consent was obtained in each case within 24 hours of the study. Parental written informed consent was obtained for a 10-year-old baby with a single liver lesion in the left lobe. Proof of FNH Burden To document proof of FNH, we reviewed medical and surgical records for all patients. Specifically, pathological proof was obtained by means of percutaneous (N ⫽ 9) or laparoscopic (N ⫽ 2) needle biopsy, including histologic core-needle and aspiration cytologic analysis in all cases. In those patients with biopsy-proven FNH, biopsy of only one nodule was performed. Considering the relatively high prevalence of multifocal FNHs within the same liver (5), patients with multiple nodules were considered to have multifocal FNH when the other lesions had the same imaging appearance of the biopsyproven FNH. Histopathologic confirmation of FNH was based on demonstration of a central fibrous region containing malformed large arteries but usually no portal veins, hyperplastic normal-appearing hepatocytes arranged in one to two cell-thick plates, associated with a prominent bile ductular reaction, and scattered inflammatory cells (24). In eight patients, biopsy was not performed because FNH lesions exhibited typical imaging features at either multiphasic CT of the liver and MRI deemed sufficient for lesion characterization (7,25). Specifically, the imaging criteria to be needed for a confident diagnosis of FNH were: a focus of homogenous bright enhancement during the hepatic arterial-dominant phase, as well as the presence of a central scar showing slight hyperintensity on T2-weighted sequences with contrast medium accumulation on the delayed phase (acquired after five minutes from the start of contrast medium injection) (26). Additional imaging features included absence of pseudocapsule surrounding the lesion on the portal venous and equilibrium phases, isointensity or slight hypointensity on precontrast T1-weighted sequences, and isointensity or slight hyperintensity on T2-weighted sequences (7). Moreover, all patients (N ⫽ 8) without pathologic confirmation of FNH underwent additional clinical and imaging follow-up, for a minimum of 12 months (mean follow-up ⫽ 24 months; follow-up range ⫽ 12–36 months), with dynamic contrastenhanced CT and MR during the hepatic arterialdominant phase (HAP, 25 seconds from the start of contrast medium administration), the portal-venous phase (PVP, 60 seconds from the start of contrast medium administration), and the equilibrium phase (EP, 180 seconds from the start of contrast medium administration). In these patients, FNH nodules (N ⫽ 8) showed size stability at imaging follow-up. Contrast Agents Ferucarbotran (Resovist威; Schering AG, Berlin, Germany) is a new SPIO composed of carboxydextrancoated iron oxide particles (magnetite ⫽ Fe3O4/ maghemite ⫽ ␥Fe2O3; relaxivity 1 ⫽ 25.4 mmol/liter–1 䡠 second–1; relaxivity 2 ⫽ 151.0 mmol/liter–1 䡠 second–1) with a hydrodynamic particle size measuring between
MR Imaging With Liver-Specific Contrast Agents
45 and 60 nm (27). The carboxydextran coating ensures aqueous solubility of the contrast agent and prevents aggregation. Ferucarbotran contains 0.5 mmol/liter of iron, 40 mg/mL of mannitol, and 2 mg/mL of lactic acid, adjusted to a pH of 6.5. At 37°C, the solution has a low viscosity (1.031 MPas) and is isotonic to blood plasma (0.319 osmol/kg H2O). After intravenous injection, about 85% of the administered ferucarbotran dose is taken up by Kupffer cells and 8% to 9% by the macrophages of the spleen. The larger particles (⬇60 nm) are taken up faster than the smaller particles (⬇45 nm) with plasma half-lives of approximately five minutes and up to ⬇100 minutes, respectively. Because the smaller particles remain in the vessels longer, ferucarbotran also has “blood-pool” characteristics. In our study, the dose of ferucarbotran was 0.9 mL for patients with a body weight less than 60 kg and 1.4 mL for patients with a body weight of 60 kg or more (medium dose ⫽ 10 mol Fe/kg; range ⫽ 7.0-12.9 mol Fe/kg equivalent to about 1% of normal whole-body iron content), which was demonstrated as effective for lesion detection and characterization at delayed imaging (16,27). To avoid patient position changes during the examination, contrast agent was prefilled in the distal part of a connecting line and administered by fast intravenous injection (flow rate ⫽ 2 mL/second) using a power injector (Spectris威; Medrad, Indianola, PA, USA) through a catheter in an antecubital vein (18 –20 gauge). Contrast medium administration was immediately followed by 20 mL saline flush (0.9%). Since the brownish color of the solution does not allow visual inspection, ferucarbotran was administered via a 5-m filter included in the commercial package. Gadobenate dimeglumine (MultiHance威; Bracco, Milan, Italy) is a sterile, intravenously injectable contrast material composed of an octadentate chelate of the paramagnetic ion, gadolinium. Gadobenate dimeglumine differs from conventional gadolinium-based agents in that it possesses a two-fold greater T1 relaxivity in human plasma (9.7 liters 䡠 mmol–1 䡠 second–1 at 0.47 T) because of weak and transient interaction of the gadolinium (Gd)benzyloxypropionictetraacetate (BOPTA) contrast-effective moiety with serum albumin (13). After an initial distribution to the extracellular fluid space, unlike other commercially available gadolinium compounds, gadobenate dimeglumine is selectively taken up by functioning hepatocytes (2– 4% of the injected dose) and excreted into the bile by the canalicular multispecific organic anion transporter (14,28). Liver-specific delayed images yield an increase in liver-lesion contrast, improving the diagnostic efficacy of gadobenate dimeglumine for liver imaging (29). Gadobenate dimeglumine was injected by bolus at a dose of 0.1 mmol/kg bodyweight (0.2 mL/kg dose of a 0.5 mol/L solution) with a 2 mL/second flow rate, followed by a 20-mL saline flush (0.9%) at the same flow rate using a power injector (Spectris威; Medrad, Indianola, PA, USA). MR Imaging Protocol All patients underwent two different MRI examinations. To minimize the risk of imaging interferences due to the prolonged T2-effect of ferucarbotran, gadobenate dime-
777
glumine-enhanced MRI was performed first. Moreover, a mean interval period of two weeks (range ⫽ 10 –20 days) was allowed between gadobenate dimeglumineand ferucarbotran-enhanced MRI. All MR studies were performed with a superconducting magnet system at a field strength of 1.5 T (Magnetom Vision; Siemens Medical System, Erlangen, Germany), with a maximum gradient field strength of 25 mT/m, and equipped with a body coil for excitation and a quadrature phased-array multicoil for signal reception. The reception coil was properly positioned to optimally cover the whole liver. In both sessions, the imaging protocol included unenhanced T2-weighted half-Fourier acquisition singleshot turbo spin-echo (HASTE) sequences (TR/TE ⫽ infinite/90 msec, echo-train length ⫽ 104, section thickness ⫽ 6 mm, intersection gap ⫽ 20%, field of view ⫽ 350 – 400 mm, effective matrix size ⫽ 192 ⫻ 256, signal average ⫽ 1, and half-Fourier reconstruction), as well as T1-weighted 2D fast low angle shot (FLASH) spoiled gradient-echo sequences (TR/TE ⫽ 160 msec/ 4.8 msec, flip angle ⫽ 80°, section thickness ⫽ 6 mm, intersection gap ⫽ 10%, field of view ⫽ 350 – 400 mm, effective matrix size ⫽ 140 ⫻ 256; signal average ⫽ 1) with and without fat saturation. Precontrast T2weighted turbo spin echo (TSE) sequences (TR/TE ⫽ 3900 msec/60 msec) with short echo train (30) were also acquired before ferucarbotran administration. Following contrast medium injection (either gadobenate dimeglumine in the first imaging session, or ferucarbotran in the second imaging session), T1-weighted fatsaturated 2D FLASH sequences were dynamically acquired during the HAP, PVP, and EP. Hepatospecific phases were acquired for gadobenate dimeglumine, using T1-weighted 2D FLASH sequences during the hepatobiliary phase (120 minutes from the start of gadobenate dimeglumine administration) (28,31), and for ferucarbotran, using T2-weighted TSE sequences during the RES-specific accumulation phase (10 minutes from the start of ferucarbotran administration). All MR sequences imaged the whole liver in the transaxial plane during a single breathhold. A rectangular field of view was used to improve spatial resolution by maximizing the number of phase-encoding lines for a given acquisition time. Presaturation bands were positioned above and below the imaging volume for T1-weighted gradient-echo sequences to minimize flow-related artifacts. Image quality was considered adequate in all cases. Image Analysis Image analysis was performed separately and independently by two abdominal radiologists (with seven and 16 years of experience, respectively) in two different reading sessions. To minimize memory bias, reading sessions were separated by a time interval of four weeks and images were presented in random order to each of the readers. The observers analyzed the gadobenate dimeglumine– enhanced MR image sets in the first reading session and the ferucarbotran-enhanced image sets in the second reading session. Both readers had knowledge of the diagnosis of FNH, and of the number
778
Marin et al.
Table 1 Morphologic Characteristics and Enhancement Pattern of 24 Proven FNHs on Dynamic T1-Weighted GRE MR Images* Signal intensitya Hyperintense Gadobenate dimeglumine Ferucarbotran Isointense Gadobenate dimeglumine Ferucarbotran Hypointense Gadobenate dimeglumine Ferucarbotran
Dynamic phase T1-weighted Arterial
Portal venous
Equilibrium
24 (100%) 2 (8%)
20 (83%) –
13 (54%) –
– 16 (67%)
4 (17%) 9 (37%)
11 (46%) 14 (58%)
– 6 (25%)
– 15 (63%)
– 10 (42%)
*Data are number and percent of lesions. Signal intensity of FNH relative to that of the surrounding liver.
a
and segmental location of liver lesions, but were blinded as to which contrast agent was used. Hepatic lesions were classified into an eight-segment partition system according to the Couinaud and Bismuth segmental anatomy of the liver (32,33). A study coordinator was present at both reading sessions, ensuring that observers assessed only the same, confirmed FNH lesions. Because we were not interested in evaluating the detection sensitivity of the two different contrast agents, when disagreement occurred, a consensus was obtained by discussion between the two radiologists. Commercially available hardware (Dell Precision 650; Dell Computer Corporation, Round Rock, TX, USA) and software (Vitrea 3; Vital Images, Minneapolis, MN, USA) were used for image analysis, which consisted of the evaluation of magnified transverse MR images directly on a high-resolution monitor. Window setting was adjusted as needed in each case. During each reading session, investigators were asked to record the size, signal intensity, and enhancement pattern of all FNH lesions. Lesion size was estimated by measuring the maximum diameter on transverse MR images with an electronic ruler. The following classification system was used to define the lesion signal intensity: isointense; hypointense; and hyperintense (34). Specifically, on precontrast T1weighted images, lesions were classified as isointense if signal intensity was similar to that of the surrounding liver, hypointense if signal intensity was similar to or lower than that of paraspinal muscles, and hyperintense if signal intensity was greater than that of the liver. On T2-weighted images, lesions were categorized as isointense if the signal intensity was similar to that of the liver, hypointense if the signal intensity was lower than that of paraspinal muscles, and hyperintense if the signal intensity was similar to the spleen. On postcontrast images (i.e., HAP, PVP, EP, and hepatospecific phase), the lesion’s signal intensity was determined by comparison to the normal surrounding liver. Finally, signal intensity was classified as slightly hypo/hyperintense if a lesion exhibited an intermediate signal intensity. Lesion enhancement was considered homogeneous if it enhanced to the same degree in all its parts, with the exception of the scar or septa. Lesion-enhancement
pattern during the hepatospecific phase was also assessed. Statistical Analysis The different percentages of hyper-, iso-, or hypointense lesions in each vascular phase, after gadobenate dimeglumine and ferucarbotran injection, respectively, were compared using McNemar’s test with Bonferroni adjustment. A P-value of 0.05 was considered statistically significant. Kappa statistics were calculated to evaluate the agreement of signal intensities of the lesions at different vascular phases with both contrast agents. The strength of agreement was assessed according to the Landis and Koch classification system: -values were judged as poor, ⬍ 0; slight, ⫽ 0.0 – 0.2; fair, ⫽ 0.21– 0.40; moderate, ⫽ 0.41– 0.60; good, ⫽ 0.61– 0.80; and excellent, ⫽ 0.81–1.00 (35). All statistical analyses were performed using commercially available software (SPSS 11.0; SPSS Inc., Chicago, IL, USA). RESULTS No adverse reactions occurred during or in the 24 hours following contrast medium infusion at either MRI session. A summary of our study results is presented in Table 1. The mean maximum lesion diameter was 4 cm (range ⫽ 1–12 cm). On precontrast T2-weighted images, 14 of 24 (58%) FNHs were isointense compared to the liver; whereas the remaining 10 (10/24; 42%) were slightly hyperintense. On precontrast T1-weighted images, 13 of 24 (54%) FNHs were isointense compared to the liver; whereas the remaining 11 (11/24; 46%) were hypointense. Dynamic T1-Weighted Phases With regard to HAP, all FNHs (24/24; 100%) showed strong and homogeneous hyperintensity on gadobenate dimeglumine-enhanced T1-weighted images (Figs. 1 and 2), whereas only two FNHs (2/24; 2%) appeared hyperintense on ferucarbotran-enhanced T1-weighted
MR Imaging With Liver-Specific Contrast Agents
images (Fig. 2e). On ferucarbotran-enhanced images during HAP, the majority of the lesions were either isointense (16/24; 67%) (Fig. 1f) or hypointense (6/24; 25%). During PVP and EP, no FNH showed hypointensity on gadobenate dimeglumine-enhanced images (Figs. 1c and d, 2c and d). By contrast, in a consistent number of cases, FNH showed a signal hypointensity relative to the liver on ferucarbotran-enhanced images on either PVP (15/24; 63%) and EP (10/24; 42%) (Figs. 1g and f, 2f and g). The results of McNemar’s test demonstrated that, during all dynamic T1-weighted phases (i.e., HAP, PVP, and EP), significantly more lesions were hyperintense on gadobenate dimeglumine-enhanced images than on ferucarbotran-enhanced images (HAP, P ⫽ 0.0000009; PVP, P ⫽ 0.0000038; EP, P ⫽ 0.0004882). By contrast, during HAP, significantly more lesions were isointense on ferucarbotran-enhanced images than on gadobenate dimeglumineenhanced images (P ⫽ 0.0000610); whereas, during PVP
779
and EP, significantly more lesions were hypointense on ferucarbotran-enhanced images than on gadobenate dimeglumine-enhanced images (P ⫽ 0.0001220 and P ⫽ 0.0039062, respectively). Overall, the kappa values calculated to evaluate the agreement of signal intensities of the lesions following administration of the two contrast agents showed poor agreement during all dynamic T1-weighted images (HAP, ⫽ – 0.5; PVP, ⫽ – 0.3; EP, ⫽ 0.04). Hepatospecific Phases On the T1-weighted gadobenate dimeglumine-enhanced hepatospecific phase, 20 FNHs showed marked and long-lasting contrast medium retention comparable or even higher than that of the normal surrounding liver, thus showing isointensity (13/24; 54%) (Fig. 1e) or hyperintensity (7/24; 29%). By contrast, four FNHs (4/24; 17%) were hypointense. On the T2-weighted ferucarbotran-enhanced hepatospecific phase, most FNHs (20/24; 83%) showed a substantial decrease of signal intensity and were isointense to the surrounding liver parenchyma because of the strong susceptibility effects of clustered intralesion SPIO particles (Fig. 1i). By contrast, four of 24 lesions
Figure 1. Typical FNH (arrows) of liver segment VII in an asymptomatic 31-year-old woman. a: Transverse noncontrast, T1-weighted, 2D fat-suppressed, FLASH spoiled gradient-echo sequence (TR/TE ⫽ 160 msec/4.8 msec; flip angle ⫽ 80°). FNH is slightly hypointense relative to the surrounding liver; a central fibrous scar (arrowhead) can also be seen. b–e: Transverse, contrast-enhanced, T1-weighted, 2D fat-suppressed FLASH spoiled gradient-echo sequences (TR/TE ⫽ 160 msec/ 4.8 msec; flip angle ⫽ 80°) acquired during the dynamic and hepatospecific phases following administration of gadobenate dimeglumine. b: HAP image after gadobenate dimeglumine administration. FNH enhances brightly and homogeneously except for the central scar and thin septa, which remain hypointense (arrowhead). c: PVP scan at the same level after gadobenate dimeglumine administration. FNH and scar begin to fade into isointensity. d: EP scan after gadobenate dimeglumine administration obtained at the same level. Lesion (arrow) is almost isointense; central scar (arrowhead) shows faint hyperintensity because of delayed contrast medium retention. e: Image obtained during the hepatospecific phase (120-minute delay) after gadobenate dimeglumine administration. FNH (arrow) appears homogeneously isointense; scar (arrowhead) is hypointense. f–h: Corresponding transverse, T1-weighted, 2D fat-suppressed, FLASH spoiled gradient-echo sequences (TR/ TE ⫽ 160 msec/4.8 msec; flip angle ⫽ 80°) acquired during the dynamic phases following administration of ferucarbotran. f: HAP image after ferucarbotran administration. FNH (arrow) is isointense to the surrounding liver parenchyma; the central fibrous scar is hypointense (arrowhead). Note high signal intensity of the aorta and patchy enhancement of the spleen (Sp), based on the vascular phase. g,h: PVP and EP images after administration of ferucarbotran. On both phases, the lesion (arrow) shows slight signal hypointensity; central scar shows no apparent contrast medium retention on EP (h). i: T2-weighted turbo spin echo (TSE) sequence (TR/TE ⫽ 3900 msec/60 msec) during the hepatospecific phase (10-minute delay) after ferucarbotran administration. FNH (arrow) appears homogeneously isointense.
780
Marin et al.
DISCUSSION
Figure 2. MR images of a large, subcapsular FNH (arrows) located in the right lobe of the liver in a 42-year-old woman with right upper quadrant pain. a: Transverse, noncontrast, T1-weighted, 2D fat-suppressed, FLASH spoiled gradient-echo sequence (TR/TE ⫽ 160 msec/4.8 msec; flip angle ⫽ 80°). FNH is slightly hypointense to the surrounding liver. Transverse, contrast-enhanced T1-weighted, 2D fat-suppressed, FLASH spoiled gradient-echo sequences (TR/TE ⫽ 160 msec/4.8 msec; flip angle ⫽ 80°) acquired after administration of gadobenate dimeglumine (b–d) and ferucarbotran (e–g). b: HAP image after gadobenate dimeglumine administration. FNH (arrow) shows intense and homogeneous enhancement except for the central scar (arrowhead) and thin septa, which remain hypointense. c,d: PVP and EP images at the same level after gadobenate dimeglumine administration. Lesion is slightly and homogeneously hyperintense; the central fibrous scar (arrowhead) remains hypointense. e: Corresponding HAP image after ferucarbotran administration. FNH shows slight homogeneous hyperintensity except for the central fibrous scar (arrowhead). f,g: PVP and EP images after administration of ferucarbotran. Lesion shows homogeneous and slight hypointensity to liver, a finding that could be misinterpreted as lesion washout.
(17%) exhibited little signal loss and were hyperintense relative to the liver. Overall, in the majority of cases (20/24; 83%), FNH showed a typical behavior on hepatospecific phase using either contrast agent, thus allowing for a more confident diagnosis.
In addition to the strong effect in shortening T2 relaxation times, SPIO compounds also have a high r1 relaxivity, which enables T1-weighted imaging with strategies similar to those used for paramagnetic lowmolecular-weighted gadolinium complexes (36,37). Due to the relatively high rate of dose-dependent side effects (i.e., acute hypotensive reactions and lumbar pain), previous studies investigated only the T2 and T2* properties of SPIO compounds during the delayed RESspecific accumulation phase. By allowing a clinically safe fast-bolus injection, ferucarbotran is the first SPIO agent that has been tested to perform first-pass dynamic T1-weighted imaging during different vascular phases, thus enabling investigation of the perfusion patterns of focal liver lesions (16 –22). At present, to our knowledge, only a few preliminary reports have been published on ferucarbotran-enhanced T1-weighted dynamic MRI (16 –22). Vogl et al (20) evaluated the diagnostic potential of ferucarbotran for a wide variety of benign and malignant liver lesions on T1-weighted dynamic MR images. This study also included 15 FNHs, but lesion enhancement patterns were not investigated. In a recent study, Reimer et al (16) described variable enhancement for hypervascular liver lesions, such as hepatocellular carcinoma (HCC) or FNH, on ferucarbotran-enhanced T1-weighted images during the early perfusion phase, but inferences could not be gathered due to the limited sample size (i.e., only one HCC and three FNHs were investigated). The results of our study show that the dynamic enhancement pattern of FNH significantly differs on ferucarbotran- and gadobenate dimeglumine– enhanced T1-weighted dynamic MR images. Kappa statistics showed total disagreement between FNH enhancement patterns during all vascular phases (i.e., HAP, PVP, and EP). In agreement with previous investigations (14,15), homogeneous and bright enhancement on HAP was the most reliable MR sign for FNH after gadobenate dimeglumine administration (24/24; 100%), whereas only two (2/24; 2%) FNHs showed positive enhancement on HAP following ferucarbotran administration. Our results differ from previous studies, in which SPIO agents induced a transient positive enhancement of the liver and spleen on T1-weighted images acquired during the first 30 seconds after the start of injection (16,17). In our study, despite the strong vascular blood supply, most FNHs showed no substantial enhancement on T1-weighted sequences acquired during HAP after ferucarbotran administration, and were isointense to the liver. Notably, this finding is in agreement with two recent studies (21,22) that intraindividually compared ferucarbotran-enhanced MRI with multidetector row CT and gadolinium-enhanced MRI in a patient population with histologically-proven HCC lesions. In both studies, HCCs showed an atypical enhancement pattern on SPIO-enhanced dynamic MR images, and, thus, these images were not judged of value in daily clinical practice for the preoperative detection of HCC because of the low contrast between hypervascular tumors and surrounding liver.
MR Imaging With Liver-Specific Contrast Agents
In contrast, on ferucarbotran-enhanced MRI, most FNHs were identified as hypointense liver nodules on PVP (15/24; 63%) and EP (10/24; 42%), whereas no FNH was hypointense on either PVP and EP after gadobenate dimeglumine administration. It should be emphasized that a persistent enhancement beyond the arterial phase has generally been considered as a useful criterion to suggest benignity of a liver lesion. The hypointense signal of FNH on ferucarbotran-enhanced MR images during PVP and EP may be mistaken for a lesion washout, that causes diagnostic difficulties in the differential diagnosis with other malignant hypervascular masses. Previous studies investigating the kinetics of ferucarbotran in animal and human subjects demonstrated that SPIO particles are progressively removed from the blood by RES cells during PVP and EP (36). As a consequence, signal intensity decrease of the liver and spleen on either T1- or T2weighted images, due to the predominant T2 effect of clustered intracellular iron oxides, and bright, persistent, blood vessel signal intensity due to the lower concentration of SPIO particles that are still circulating, are generally observed. However, reasons for the higher signal drop (i.e., hypointensity) of most FNHs relative to the liver parenchyma still remain unclear. Further studies are needed to specifically investigate this issue. Moreover, eight FNHs (8/24; 33%) could not be detected on ferucarbotran-enhanced dynamic MRI, due to signal isointensity to the liver during all vascular phases. The different enhancement patterns of FNH following administration of gadobenate dimeglumine and ferucarbotran can be explained by several factors. First, as recently supposed by Lutz et al (22), both contrast agents have different distribution volumes. Low-molecular gadolinium chelates, such as gadobenate dimeglumine, show rapid extravasation into the interstitium. For this contrast agent, both the vascular and interstitial effects contribute equally to lesion enhancement. Conversely, due to the longer blood half-life of SPIO compounds, intravascular concentrations of ferucarbotran minimally change during perfusion studies (the so-called “blood pool effect”) (36). Second, in our study, the two contrast agents were administered at different doses. Indeed, ferucarbotran was injected using the recommended dose to perform optimal T2-weighted imaging (medium dose ⫽ 10 mol Fe/kg) (27), whereas a 10-fold higher concentration of gadobenate dimeglumine (medium dose ⫽ 0.1 mmol/kg) was routinely administered in the same patient. However, it should also be emphasized that, within the range of diagnostically applied proton Larmor frequencies, the r1 relaxivity of ferucarbotran is four times higher than the r1 of low molecular weight gadolinium compounds (16). Moreover, due to the strong r2 relaxivity, the administration of SPIO agents at doses higher than 10 mol Fe/kg has been associated with negative enhancement of the liver on T1-weighted images (16,37). A third variable should be considered to explain our results with regard to the degree of T1 weighting. We performed a standard protocol for both MR examinations using T1-weighted images obtained at the shortest in-phase TE (4.8 msec). However, as indicated by Reimer et al (16), it can be hypothesized that shorter
781
echo times would further reduce the T2 effect of ferucarbotran, thus enabling the clinical use of more sensitive T1-weighted images. With regard to the hepatospecific phases, our results differ from those reported in a recent paper by Grazioli et al (15). Specifically, in our series, a typical enhancement pattern was noticed for an equal number of FNHs (20/24; 83%) on either gadobenate dimeglumine– and ferucarbotran-enhanced MR images. Although we have no definite explanations for such discrepancies, it may be attributed to the different SPIO agent investigated and, at least to a certain extent, to the shorter T2-weighting of our T2weighted TSE sequences during the RES-specific accumulation phase (30). In agreement with previous investigations (13–15), most FNHs were isointense (13/24; 54%) or hyperintense (7/24; 29%) to the liver during the hepatospecific phase after gadobenate dimeglumine administration. These findings typically reflect the ability of the functioning hepatocytes contained in FNH to take up the contrast medium across the sinusoidal membrane and excrete it into the malformed blind-ending primary bile ductules, thus resulting in a persistent enhancement of the lesion (14). On the T2-weighted ferucarbotran-enhanced hepatospecific phase, most FNHs (20/24; 83%) showed a signal intensity decrease comparable to that of the normal surrounding liver and were isointense. This is due to the strong susceptibility effects of clustered SPIO particles in Kupffer cells, which are normally present in FNH. In contrast, in an equal number of cases (4/24; 17%), the hepatospecific phase showed atypical signal intensities following either gadobenate dimeglumine or ferucarbotran administration and, therefore, it was not deemed useful for lesion characterization. As previously asserted by some authors (15,38), a comparatively small number of Kupffer cells may be responsible for the insufficient signal drop of FNH occasionally seen after administration of SPIO compounds. By contrast, we have no persuasive explanations concerning the limited number of lesions lacking gadobenate dimeglumine accumulation during the liver-specific hepatobiliary phase. In general, however, it should be emphasized that, in clinical practice, hepatospecific phases usually have limited value for FNH detection for both contrast agents, due to low lesion-to-liver contrast. Our study has several limitations. First, there is a selection bias because all recruited patients had a definite diagnosis of FNH. Moreover, the number of examined lesions was relatively small. However, a lesion-bylesion comparison of both contrast agents was undertaken in the same patients. Another limitation of our study is the lack of histopathologic proof for every lesion in every patient. Indeed, in eight patients with typical imaging features of FNH, lesions were not biopsied and the benign nature of the lesion was clinically confirmed by stable behavior at imaging follow-up. Furthermore, when multiple lesions were present, biopsy was performed on one nodule only; other lesions were presumed to be FNH if they exhibited imaging features similar to the biopsy-proven nodule. However, within the limits of ethical patient care, we believe that typical imaging features and follow-up deemed sufficient evidence of the benign nature and likely histologic characteristics in all nonbiopsied lesions.
782
Finally, we did not investigate the diagnostic value of ferucarbotran-enhanced T1-weighted images in the hepatospecific phase (10 minutes) for FNH detection and characterization (39). In conclusion, the enhancement patterns of FNH on T1-weighted dynamic perfusion imaging did not correlate after ferucarbotran and gadobenate dimeglumine administration. Specifically, when both contrast agents were directly compared in different vascular phases within the same lesion, the well-recognized marked and homogeneous lesion-enhancement of FNH on HAP at gadolinium-enhanced dynamic MRI was an uncommon finding after ferucarbotran administration. Further studies are needed to achieve a valuable dynamic phase after fast bolus injection of ferucarbotran, as dynamic T1-weighted imaging after gadolinium injection is still the cornerstone for FNH identification and characterization. By contrast, on hepatospecific phases, the majority of FNHs showed a typical behavior on either gadobenate dimeglumine and ferucarbotran MRI, thus allowing for a more confident diagnosis.
Marin et al.
16.
17.
18.
19.
20.
21.
22.
23.
REFERENCES 1. Wanless IR, Mawdsley C, Adams R. On the pathogenesis of focal nodular hyperplasia of the liver. Hepatology 1985;5:1194 – 1200. 2. Becker YT, Raiford DS, Webb L, Wright JK, Chapman WC, Pinson CW. Rupture and hemorrhage of hepatic focal nodular hyperplasia. Am Surg 1995;61:210 –214. 3. Charny CK, Jarnagin WR, Schwartz LH, et al. Management of 155 patients with benign liver tumours. Br J Surg 2001;88:808 – 813. 4. Reddy KR, Kligerman S, Levi J, et al. Benign and solid tumors of the liver: relationship to sex, age, size of tumors, and outcome. Am Surg 2001;67:173–178. 5. Brancatelli G, Federle MP, Grazioli L, Blachar A, Peterson MS, Thaete L. Focal nodular hyperplasia: CT findings with emphasis on multiphasic helical CT in 78 patients. Radiology 2001;219:61– 68. 6. Ferlicot S, Kobeiter H, Tran Van Nhieu J, et al. MRI of atypical focal nodular hyperplasia of the liver: radiology-pathology correlation. AJR Am J Roentgenol 2004;182:1227–1231. 7. Choi CS, Freeny PC. Triphasic helical CT of hepatic focal nodular hyperplasia: incidence of atypical findings. AJR Am J Roentgenol 1998;170:391–395. 8. Ba-Ssalamah A, Schima W, Schmook MT, et al. Atypical focal nodular hyperplasia of the liver: imaging features of non-specific and liver-specific MR contrast agents. AJR Am J Roentgenol 2002;179: 1447–1456. 9. Attal P, Vilgrain V, Brancatelli G, et al. Telangiectatic focal nodular hyperplasia: US, CT, and MR imaging findings with histopathologic correlation in 13 cases. Radiology 2003;228:465– 472. 10. Vilgrain V, Uzan F, Brancatelli G, Federle MP, Zappa M, Menu Y. Prevalence of hepatic hemangioma in patients with focal nodular hyperplasia: MR imaging analysis. Radiology 2003;229:75–79. 11. Friedman LS, Gang DL, Hedberg SE, Isselbacher KJ. Simultaneous occurrence of hepatic adenoma and focal nodular hyperplasia: report of a case and review of the literature. Hepatology 1984;4:536 –540. 12. Spinazzi A, Lorusso V, Pirovano G, Kirchin M. Safety, tolerance, biodistribution, and MR imaging enhancement of the liver with gadobenate dimeglumine: results of clinical pharmacologic and pilot imaging studies in nonpatient and patient volunteers. Acad Radiol 1999;6:282–291. 13. Grazioli L, Morana G, Kirchin MA, Schneider G. Accurate differentiation of focal nodular hyperplasia from hepatic adenoma at gadobenate dimeglumine– enhanced MR imaging: prospective study. Radiology 2005;236:166 –177. 14. Grazioli L, Morana G, Federle MP, et al. Focal nodular hyperplasia: morphologic and functional information from MR imaging with gadobenate dimeglumine. Radiology 2001;221:731–739. 15. Grazioli L, Morana G, Kirchin MA, et al. MRI of focal nodular hyperplasia (FNH) with gadobenate dimeglumine (Gd-BOPTA) and
24.
25. 26.
27. 28.
29.
30.
31.
32. 33. 34.
35. 36.
37.
38.
39.
SPIO (ferumoxides): an intra-individual comparison. J Magn Reson Imaging 2003;17:593– 602. Reimer P, Muller M, Marx C, et al. T1 effects of a bolus-injectable superparamagnetic iron oxide, SH U 555 A: dependence on field strength and plasma concentration—preliminary clinical experience with dynamic T1-weighted MR imaging. Radiology 1998;209: 831– 836. Chen F, Ward J, Robinson PJ. MR imaging of the liver and spleen: a comparison of the effects on signal intensity of two superparamagnetic iron oxide agents. Magn Reson Imaging 1999;17:549 –556. Muller M, Reimer P, Wiedermann D, et al. [T1-weighted dynamic MRI with new superparamagnetic iron oxide particles (Resovist): results of a phantom study as well as 25 patients]. Rofo 1998;168: 228 –236. [German] Weishaupt D, Willmann JK, Lutz AM, Marincek B. Resovist for imaging of hepatocellular carcinoma in the cirrhotic liver. Eur Radiol 2004;14(Suppl 1):C5–C6. Vogl TJ, Schwarz W, Marquart F, et al. Experience with Resovist in gastroenterologic radiology and intervention. Eur Radiol 2004; 14(Suppl 1):C7–C9. Kim SH, Choi D, Kim SH, et al. Ferucarbotran-enhanced MRI versus triple-phase MDCT for the preoperative detection of hepatocellular carcinoma. AJR Am J Roentgenol 2005;184:1069 –1076. Lutz AM, Willmann JK, Goepfert K, Marincek B, Weishaupt D. Hepatocellular carcinoma in cirrhosis: enhancement patterns at dynamic gadolinium- and superparamagnetic iron oxide-enhanced T1-weighted MR imaging. Radiology 2005;237:520 –528. World Medical Association Declaration of Helsinki. Recommendations guiding physicians in biomedical research involving human subjects. JAMA 1997;277:925–956. Nguyen BN, Flejou JF, Terris B, Belghiti J, Degott C. Focal nodular hyperplasia of the liver: a comprehensive pathologic study of 305 lesions and recognition of new histologic forms. Am J Surg Pathol 1999;23:1441–1454. Carlson SK, Johnson CD, Bender CE, Welch TJ. CT of focal nodular hyperplasia of the liver. AJR Am J Roentgenol 2000;174:705–712. Vilgrain V, Flejou JF, Arrive L, et al. Focal nodular hyperplasia of the liver: MR imaging and pathologic correlation in 37 patients. Radiology 1992;184:699 –703. Reimer P, Rummeny EJ, Daldrup HE, et al. Clinical results with Resovist: a phase 2 clinical trial. Radiology 1995;195:489 – 496. Spinazzi A, Lorusso V, Pirovano G, Taroni P, Kirchin M, Davies A. Multihance clinical pharmacology: biodistribution and MR enhancement of the liver. Acad Radiol 1998;5(Suppl 1):S86 –S89; discussion S93–S94. Petersein J, Spinazzi A, Giovagnoni A, et al. Focal liver lesions: evaluation of the efficacy of gadobenate dimeglumine in MR imaging—a multicenter phase III clinical study. Radiology 2000;215:727–736. Frohlich JM. MRI of focal nodular hyperplasia (FNH) with gadobenate dimeglumine (Gd-BOPTA) and SPIO (ferumoxides): an intraindividual comparison. J Magn Reson Imaging 2004;19:375–376. Vogl TJ, Pegios W, McMahon C, et al. Gadobenate dimeglumine-a new contrast agent for MR imaging: preliminary evaluation in healthy volunteers. AJR Am J Roentgenol 1992;158:887– 892. Couinaud C. Le foie: e´tude anatomiques et chirurgicales. Paris: Masson; 1957. p 9 –12. Bismuth H. Surgical anatomy and anatomical surgery of the liver. World J Surg 1982;6:3–9. Mastropasqua M, Braga L, Kanematsu M, et al. Hepatic nodules in liver transplantation candidates: MR imaging and underlying hepatic disease. Magn Reson Imaging 2005;23:557–562. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159 –174. Frank H, Weissleder R, Brady TJ. Enhancement of MR angiography with iron oxide: preliminary studies in whole-blood phantom and in animals. AJR Am J Roentgenol 1994;162:209 –213. Oswald P, Clement O, Chambon C, Schouman-Claeys E, Frija G. Liver positive enhancement after injection of superparamagnetic nanoparticles: respective role of circulating and uptaken particles. Magn Reson Imaging 1997;15:1025–1031. Paley MR, Mergo PJ, Torres GM, Ros PR. Characterization of focal hepatic lesions with ferumoxides-enhanced T2-weighted MR imaging. AJR Am J Roentgenol 2000;175:159 –163. Oudkerk M, van den Heuvel AG, Wielopolski PA, Schmitz PI, Borel Rinkes IH, Wiggers T. Hepatic lesions: detection with ferumoxide-enhanced T1-weighted MR imaging. Radiology 1997; 203:449 – 456.
