HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2342040278v1 234/2/468    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Huppertz, A. Articles by Reiser, M. F. Search for Related Content PubMed PubMed Citation Articles by Huppertz, A. Articles by Reiser, M. F.

Enhancement of Focal Liver Lesions at Gadoxetic Acid–enhanced MR Imaging: Correlation with Histopathologic Findings and Spiral CT—Initial Observations¹

¹ From the Department of Clinical Radiology (A.H., A.K., C.J.Z., J.S., T.K.H., M.F.R.) and Institute of Pathology (S.H.), Ludwig-Maximilians University, Munich, Germany; and Department of Corporate Clinical Development Diagnostics, Schering, Berlin, Germany (A.H., J.B.). Received February 12, 2004; revision requested April 20; revision received May 6; accepted June 15. Address correspondence to A.H., Imaging Science Institute Charité-Siemens, Robert-Koch-Platz 7, 10115 Berlin, Germany (e-mail: alexander.huppertz@siemens.com).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To detect hepatocyte-selective enhancement of focal lesions with gadoxetic acid at magnetic resonance (MR) imaging and to correlate enhancement in hepatocyte-selective phases with histopathologic findings and in arterial and portal venous phases with biphasic computed tomographic (CT) findings.

MATERIALS AND METHODS: Study was supported by local ethics committee; all patients gave written informed consent. In 19 men and 14 women recruited in three clinical studies, histopathologic correlation and CT scans of 41 focal lesions (13 primary malignant lesions, 21 metastases, three adenomas, three cases of focal nodular hyperplasia [FNH], and one cystadenoma) and ultrasonographic confirmation of five cysts were available. MR was performed before and during arterial and portal venous phases and in hepatocyte-selective phases 10 and 20 minutes after injection of gadoxetic acid. Enhancement was evaluated in consensus by two observers. Enhancement pattern and morphologic features during arterial and portal venous phases were correlated between gadoxetic acid–enhanced MR and CT images by means of adjusted ² test.

RESULTS: Hepatocyte-selective uptake was observed 10 and 20 minutes after injection in FNH (three of three), adenoma (two of three), cystadenoma (one of one), and highly differentiated hepatocellular carcinoma (HCC [grade G1], two of four). Uptake was not detected in metastases (21 of 21), cholangiocarcinoma (three of three), combined hepatocellular cholangiocarcinoma (one of one), undifferentiated carcinoma (one of one), moderately or poorly differentiated HCC (grade G2–G3) (four of four), HCC (grade G1, two of four), adenoma with atypia (one of three), or cysts (five of five). During arterial and portal venous phases, there was high overall agreement rate of 0.963 between gadoxetic acid–enhanced MR and CT (simultaneous 95% confidence interval: 0.945, 0.981).

CONCLUSION: Liver-specific enhancement of focal lesions is hepatocyte selective and correlates with various histopathologic diagnoses regarding presence of certain hepatocytic functions. Arterial and portal venous MR images obtained with gadoxetic acid are comparable to those of CT.

© RSNA, 2004

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
During the past decade, several liver-specific magnetic resonance (MR) imaging contrast media have been developed and investigated in clinical studies with the objective of increasing the performance of liver MR imaging, especially for lesion detection (1). Two classes of agents can be differentiated. First, there is the class of superparamagnetic iron oxides, with the two products ferumoxides (AMI 25, Ferridex, Berlex Laboratories, Montville, NJ; and Endorem, Guerbet, Aulnay Sous Bois, France) and ferucarbotran (Resovist; Schering, Berlin, Germany), both targeted to the reticuloendothelial system in the liver. Second, there are the agents targeted to hepatocytes: gadobenate dimeglumine (Multihance; Bracco Imaging, Milan, Italy), gadoxetic acid (Primovist; Schering), and mangafodipir trisodium (Teslascan; Amersham Health, Oslo, Norway).

Gadoxetic acid is a more recently developed liver-specific MR imaging contrast agent with combined perfusion and hepatocyte-selective properties (2,3). The agent has been shown to be highly liver specific, with an uptake of about 50% by the organic anion-transporting polypeptide 1 (4,5). In clinical trials, the agent has been demonstrated to increase the detection of focal liver lesions (6) and to provide differential diagnostic information comparable to nonspecific extracellular gadolinium chelates (7,8).

In the evaluation of the clinical benefit of gadoxetic acid for lesion characterization, an independent evaluation of perfusion and hepatocyte-selective properties is necessary to analyze the contributions to the diagnosis. The reference standard for the evaluation of the arterial and portal venous perfusion phases is multiphasic spiral computed tomography (CT) or contrast material–enhanced MR imaging, which both involve the use of nonspecific extracellular agents. The characterization of lesions in nonspecific "extracellular" imaging is mostly based on radiomorphologic features (morphology at CT and MR imaging) and the particular enhancement pattern. The reference standard for the analysis of the hepatocyte-selective properties is histopathologic examination. It has to be proved whether enhancement in the liver-specific phase is hepatocyte selective and can be used for information on a cellular level.

The purpose of our study was to detect hepatocyte-selective enhancement of focal lesions with gadoxetic acid at MR imaging and to correlate enhancement in the hepatocyte-selective phases with histopathologic findings and in arterial and portal venous phases with CT findings.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The study was supported by Schering, Berlin, Germany, which provided the MR imaging contrast agent. All authors who were not employees of the company (S.H., A.K., C.J.Z., J.S., T.K.H., M.F.R.) had control of the data and the information for publication.

Study Population
Thirty-three patients recruited in three clinical studies had validation of their focal liver lesions by means of histopathologic examination and a dual-phase CT examination within a period of 6 weeks prior to or after MR imaging. They received a 0.025 mmol per kilogram of body weight dose of the liver-specific contrast medium gadoxetic acid (volume range, 4.4–10.5 mL). Of these 33 patients, 31 were examined in phase III studies, and two were examined in phase II studies. Fifteen of the phase III study patients were included in a prior publication (6) on the evaluation of detection of focal lesions. All three of the studies were approved by the local ethics committee, and all patients gave their written informed consent, which also included subsequent analysis of images and findings. The committee was informed of and approved the subgroup analysis of these studies.

Nineteen men with an age range of 31–81 years (mean age, 62.4 years) and 14 women with an age range of 35–76 years (mean age, 59.1 years) were included in the study. Statistical analysis (t test) did not show a statistically significant difference in age distribution according to patient sex.

These 33 patients (mean age, 61.0 years; mean weight, 74.2 kg; weight range, 44–105 kg) had a total of 61 lesions at CT. Five patients had liver cirrhosis (four patients had Child-Pugh class A disease, and one patient had Child-Pugh class B disease). Forty-six of these lesions were proved at histopathologic examination or transabdominal ultrasonography (US) and were defined as the set of lesions correlated in our study with the hepatocyte-specific phases and the arterial and portal venous phases.

Forty-one of these lesions were proved at histopathologic examination, and the five cysts were proved at US. In the 15 lesions detected at CT that did not have histopathologic correlation, the respective liver segment was not resected and biopsy results were not available in 14 lesions, and one lesion that appeared to be an independent lesion at CT proved to be part of the larger adjacent lesion.

Twenty-eight patients underwent surgery. Pathologic specimens revealed 35 solid lesions; in five patients, core-needle biopsy of five lesions was performed. In one of these patients, both surgery and biopsy of different lesions were performed.

One patient (62 years old, pathologic diagnosis of cholangiocellular carcinoma [CCC]) developed a severe cardiovascular and respiratory insufficiency consistent with pulmonary embolism approximately 15 hours after injection of contrast medium and died 9 days later. The causality was not considered to be related to the contrast medium. In this patient, an autopsy specimen was available to confirm the underlying disease as the cause of the event (portal vein thrombosis, femoral vein thrombosis, and pulmonary embolism).

The three phase II and III studies from which our study population of 33 patients was derived were part of multicenter trials. In total, 63 patients suspected of having or known to have focal liver lesions were enrolled at our institution between October 1994 and December 2000. Main exclusion criteria were patients younger than 18 years of age, patients treated previously with gadoxetic acid or any other investigational products within 30 days prior to study entry, patients treated with other contrast media within 24 hours prior to or after administration of the study medication, patients who had received any liver-specific agent within 2 weeks prior to gadoxetic acid administration, pregnant or nursing women, clinically unstable patients or patients who were scheduled for biopsy or liver surgery within 24 hours after administration of the study medication, and patients with history of anaphylactoid or anaphylactic reaction to any medication, including contrast media.

In the collective of 63 patients included in all three of the clinical studies, 30 were excluded from the present report because histopathologic examination was not performed (27 of 63 patients), specimens did not contain focal liver lesions (two of 63), no contrast-enhanced MR images were available because of withdrawn consent (one of 63), or a dose lower (0.0125 mmol/kg) or higher (0.05 mmol/kg) than the clinical dose was used in patients examined during the phase II study (15 of 23 patients).

Lesion Tracking
In patients who underwent surgery, consistency between CT, MR imaging, US, and pathologic findings was reached by means of a lesion-tracking procedure. For each image assessment, liver maps were completed by drawing each individual liver lesion on a respective map (according to the Couinaud system of liver anatomy) (9). This was to be done as accurately as possible (relative to lesion size and location) by one investigator (A.H.) for MR and CT images and by a different investigator (S.H.) for pathologic findings. In each assessment, a continuous number was allocated to each lesion. The surgeon did not record any individual lesions on the map but marked the areas of the liver that were resected. In patients that underwent surgery, a distinct marking of the specimen at the borders was performed to enable pathologic-anatomic correlation.

Histopathologic Procedure
Both the biopsy specimens and the resected specimens were formalin fixed, but only the latter were sliced by the pathologist in the same orientation (transverse) and with the same slice thickness used for MR imaging and CT (5–8 mm). From each of the nodular lesions, representative samples were obtained and were paraffin embedded for routine histopathologic examination. The original specimens were reviewed by a hepatic pathologist (S.H., 12 years of experience in hepatic pathology), who was blinded to the MR imaging and CT findings.

Hepatocellular carcinomas (HCCs) were classified according to the definition of the World Health Organization (high differentiation, G1; moderate differentiation, G2; and poor differentiation, G3) (10) on the basis of the definitions of Edmondson and Steiner (11). If a carcinoma contained unequivocal elements of both HCC and CCC that were intimately admixed, thereby excluding collision tumors, the lesion was classified as combined hepatocellular cholangiocarcinoma (CHCC-CC) according to the most recent Armed Forces Institute of Pathology, or AFIP, definitions of liver tumors (12). Undifferentiated carcinoma according to the World Health Organization and AFIP fits neither in the HCC nor in the CCC diagnostic category (10).

This retrospective pathologic evaluation was conducted at a higher level of differentiation compared with the first protocol-compliant evaluation in clinical practice. Thus, patients with CHCC-CC or undifferentiated carcinoma were all classified as having HCC in the respective multicenter trials and were treated with current standard care.

Contrast Agent
Gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (gadoxetic acid disodium, C₂₃H₂₈GdN₃Na₂O₁₁, molecular weight of 725.71 Da) was administered manually as a bolus through an intravenous catheter placed in the cubital vein at a speed of about 2 mL/sec. The catheter was flushed with 30 mL of 0.9% saline. Gadoxetic acid shows hepatocellular uptake via the organic anion-transporting polypeptide 1, a facilitated-type non–sodium-dependent membrane transport system also known to transport bromsulphtalein (13,14).

The biliary excretion is performed by canalicular multispecific organic anion transporter (15). Gadoxetic acid exhibits a higher T1 relaxivity in human plasma (8.7 L · mmol^–1 · sec^–1 in H₂O at 0.47 T) compared with extracellular contrast media, such as gadopentetate dimeglumine (Magnevist, Schering) (4.1 L · mmol^–1 · sec^–1), because of the higher degree of protein binding (10.0% ± 1.9) than that of gadopentetate dimeglumine (1.0% ± 1.7) (3). In human dose-finding studies, 0.025 mmol/kg has proven to be the lowest dose of gadoxetic acid that enables improved detection of hepatic lesions over gadopentetate dimeglumine–enhanced MR imaging while providing comparable differential diagnostic information (7,16). Therefore, a dose of 0.025 mmol/kg was used in clinical phase III studies (6).

Imaging Protocols
All patients were studied with a 1.5-T MR system (Magnetom Vision; Siemens, Erlangen, Germany) with high-performance gradients (25-mT/m maximum gradient strength and 600-µsec rise time) and a torso phased-array coil. T1-weighted fast low-angle shot gradient-recalled-echo (GRE) MR sequences were performed, and subsequent images were used for evaluation. Arterial and portal venous phase MR images were acquired 20 and 60 seconds after contrast material administration, and hepatocyte-selective images were acquired 10 and 20 minutes after contrast material administration (Table 1).

In the first session, MR images obtained in the hepatocyte-selective phases 10 and 20 minutes after contrast material injection were evaluated in combination with the precontrast T1-weighted MR images obtained with chemically selective fat suppression. The following liver-specific enhancement features were collected: visual increase of signal intensity (yes or no), appearance of enhancement (homogeneous, heterogeneous, or septal), and signal intensity characteristics of the lesion (compared with perceived normal parenchyma, the lesion is isointense, hypointense, or hyperintense).

In a second session, arterial dominant and portal venous phase CT and MR images were analyzed. CT scans were evaluated independently from MR images by the same readers with an interval of at least 3 weeks. The following seven radiomorphologic features were assessed at MR imaging and CT: margins (well or ill defined), presence of a capsule (no or yes; if yes, complete or incomplete), lesion appearance (homogeneous, heterogeneous, or other), presence of a scar (yes or no), evidence of intratumoral necrosis (yes or no), shape (oval, round, lobulated, or other), and presence of mass effect around the lesion (none, vessel, bile ducts, both, or other).

In addition, the following three enhancement features were evaluated for arterial and portal venous phases: enhancement (none, partial, or complete), pattern of enhancement (in the rim, homogeneous, heterogeneous, or not applicable), and MR signal intensity characteristics and CT attenuation of the lesion (compared with perceived normal parenchyma, the lesion is isointense [isoattenuating at CT], hypointense [hypoattenuating at CT], hyperintense [hyperattenuating at CT], or a combination of signal intensity characteristics).

Statistical Analysis
The agreement rate between gadoxetic acid–enhanced MR imaging and CT in arterial dominant and portal venous phases was estimated, along with confidence intervals based on an adjusted ² test (17). A two-sided significance level of 5% was used for all analyses. To evaluate the agreement of MR imaging and CT results, the difference in outcomes was calculated, with 1 for agreement and 0 for disagreement, and averaged across features and lesions by taking the correlation between features within a lesion into account. The correlation between lesions within a patient was not considered, since in most patients, only one lesion with histopathologic findings was available (22 of 33 patients).

To judge the clinical relevance of the agreement, analogous to the intraclass correlation coefficient, an agreement of at least 0.75 (lower limit of the 95% confidence interval for the agreement rate) is suggested to be meaningful and of good reliability. Calculations were performed by using SAS version 8.2 software (SAS Institute, Cary, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Histopathologic Classification of Lesions
Histopathologic evaluation revealed 34 malignant and seven benign or tumorlike lesions. The group of malignant lesions included 21 metastases (including 19 hypovascular metastases [14 of colorectal cancer, two of squamous hypopharynx carcinoma, one of thyroid carcinoma, one of endometrium adenocarcinoma, and one of adenocarcinoma of unknown primary origin] and two hypervascular metastases [one of pancreatic neuroendocrine carcinoma and one of gastric leiomyosarcoma]), eight HCCs, three CCCs, one CHCC-CC, and one undifferentiated carcinoma. In the group of benign or tumorlike lesions, three cases of focal nodular hyperplasia (FNH), two adenomas, one adenoma with atypia, and one biliary cystadenoma were detected. In addition, five cysts were confirmed with US. Overall, 21 nonhepatocellular lesions (metastases), 17 hepatocellular lesions (HCC, CCC, CHCC-CC, undifferentiated carcinoma, adenoma, and biliary cystadenoma), and eight tumorlike lesions (FNH, cysts) were identified.

Correlation of MR and CT in Arterial and Portal Venous Phases
During arterial dominant and portal venous phases, the overall agreement rate between gadoxetic acid–enhanced MR imaging and CT was 0.963 (simultaneous 95% confidence interval: 0.945, 0.981). Since the lower confidence limit for the agreement rate was above 0.75, statistical significance for a high agreement between gadoxetic acid–enhanced MR imaging and CT was found. In the evaluation of the radiomorphologic features, an agreement rate of 0.977 was found; for the enhancement features, the agreement rate was 0.947. The agreement rates for all features are listed separately in Table 3.

Differences between MR imaging and CT were observed for one CCC, a hypovascular metastasis, and the biliary cystadenoma, which all had slight peripheral or septal arterial enhancement at CT not detectable with MR imaging. Another CCC had peripheral arterial enhancement at MR imaging that was not visible at CT. In the portal venous phase, the enhancement in the neuroendocrine metastasis was washed out at MR imaging, whereas it was still present at CT. The intensity of enhancement differed in some lesions between MR imaging and CT, so that—for example—a hyperattenuating lesion at CT appeared isointense at MR imaging (Fig 1).

View larger version (129K): [in this window] [in a new window] [Download PPT slide]   Figure 1a. Metastasis of neuroendocrine tumor without hepatocyte-selective enhancement after gadoxetic acid injection in 67-year-old man with metastasis of neuroendocrine carcinoma of pancreas. Transverse MR images acquired at 1.5 T (a) before contrast material injection and (b) during arterial dominant phase. (c) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (d) Transverse MR image obtained 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Both MR image (b, T1-weighted GRE sequence without fat suppression [141.8/4.1; flip angle, 75°]) and CT scan (c) show round well-delineated lesion (arrow) with homogeneous enhancement in arterial phase. At gadoxetic acid-enhanced MR imaging, enhancement during arterial phase is seen as change from hypointensity on precontrast image (a) to isointensity during arterial phase (b). At arterial phase CT (c), lesion is hyperattenuating in comparison to surrounding parenchyma. In hepatocyte-selective phase 20 minutes after injection (d), T1-weighted GRE MR sequence with fat suppression (124.4/4.8; flip angle, 75°) shows no lesion enhancement. (e) At photomicrography of resected specimen, there is a highly vascularized highly differentiated trabecular carcinoma () infiltrating otherwise normal liver parenchyma. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (153K): [in this window] [in a new window] [Download PPT slide]   Figure 1b. Metastasis of neuroendocrine tumor without hepatocyte-selective enhancement after gadoxetic acid injection in 67-year-old man with metastasis of neuroendocrine carcinoma of pancreas. Transverse MR images acquired at 1.5 T (a) before contrast material injection and (b) during arterial dominant phase. (c) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (d) Transverse MR image obtained 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Both MR image (b, T1-weighted GRE sequence without fat suppression [141.8/4.1; flip angle, 75°]) and CT scan (c) show round well-delineated lesion (arrow) with homogeneous enhancement in arterial phase. At gadoxetic acid-enhanced MR imaging, enhancement during arterial phase is seen as change from hypointensity on precontrast image (a) to isointensity during arterial phase (b). At arterial phase CT (c), lesion is hyperattenuating in comparison to surrounding parenchyma. In hepatocyte-selective phase 20 minutes after injection (d), T1-weighted GRE MR sequence with fat suppression (124.4/4.8; flip angle, 75°) shows no lesion enhancement. (e) At photomicrography of resected specimen, there is a highly vascularized highly differentiated trabecular carcinoma () infiltrating otherwise normal liver parenchyma. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (98K): [in this window] [in a new window] [Download PPT slide]   Figure 1c. Metastasis of neuroendocrine tumor without hepatocyte-selective enhancement after gadoxetic acid injection in 67-year-old man with metastasis of neuroendocrine carcinoma of pancreas. Transverse MR images acquired at 1.5 T (a) before contrast material injection and (b) during arterial dominant phase. (c) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (d) Transverse MR image obtained 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Both MR image (b, T1-weighted GRE sequence without fat suppression [141.8/4.1; flip angle, 75°]) and CT scan (c) show round well-delineated lesion (arrow) with homogeneous enhancement in arterial phase. At gadoxetic acid-enhanced MR imaging, enhancement during arterial phase is seen as change from hypointensity on precontrast image (a) to isointensity during arterial phase (b). At arterial phase CT (c), lesion is hyperattenuating in comparison to surrounding parenchyma. In hepatocyte-selective phase 20 minutes after injection (d), T1-weighted GRE MR sequence with fat suppression (124.4/4.8; flip angle, 75°) shows no lesion enhancement. (e) At photomicrography of resected specimen, there is a highly vascularized highly differentiated trabecular carcinoma () infiltrating otherwise normal liver parenchyma. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (136K): [in this window] [in a new window] [Download PPT slide]   Figure 1d. Metastasis of neuroendocrine tumor without hepatocyte-selective enhancement after gadoxetic acid injection in 67-year-old man with metastasis of neuroendocrine carcinoma of pancreas. Transverse MR images acquired at 1.5 T (a) before contrast material injection and (b) during arterial dominant phase. (c) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (d) Transverse MR image obtained 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Both MR image (b, T1-weighted GRE sequence without fat suppression [141.8/4.1; flip angle, 75°]) and CT scan (c) show round well-delineated lesion (arrow) with homogeneous enhancement in arterial phase. At gadoxetic acid-enhanced MR imaging, enhancement during arterial phase is seen as change from hypointensity on precontrast image (a) to isointensity during arterial phase (b). At arterial phase CT (c), lesion is hyperattenuating in comparison to surrounding parenchyma. In hepatocyte-selective phase 20 minutes after injection (d), T1-weighted GRE MR sequence with fat suppression (124.4/4.8; flip angle, 75°) shows no lesion enhancement. (e) At photomicrography of resected specimen, there is a highly vascularized highly differentiated trabecular carcinoma () infiltrating otherwise normal liver parenchyma. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (190K): [in this window] [in a new window] [Download PPT slide]   Figure 1e. Metastasis of neuroendocrine tumor without hepatocyte-selective enhancement after gadoxetic acid injection in 67-year-old man with metastasis of neuroendocrine carcinoma of pancreas. Transverse MR images acquired at 1.5 T (a) before contrast material injection and (b) during arterial dominant phase. (c) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (d) Transverse MR image obtained 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Both MR image (b, T1-weighted GRE sequence without fat suppression [141.8/4.1; flip angle, 75°]) and CT scan (c) show round well-delineated lesion (arrow) with homogeneous enhancement in arterial phase. At gadoxetic acid-enhanced MR imaging, enhancement during arterial phase is seen as change from hypointensity on precontrast image (a) to isointensity during arterial phase (b). At arterial phase CT (c), lesion is hyperattenuating in comparison to surrounding parenchyma. In hepatocyte-selective phase 20 minutes after injection (d), T1-weighted GRE MR sequence with fat suppression (124.4/4.8; flip angle, 75°) shows no lesion enhancement. (e) At photomicrography of resected specimen, there is a highly vascularized highly differentiated trabecular carcinoma () infiltrating otherwise normal liver parenchyma. (Hematoxylin-eosin stain; original magnification, x200.)

Enhancement in Hepatocyte-selective Phases at MR Imaging
In the evaluation of hepatocyte-selective MR images (Table 6), absence of enhancement was observed in all 21 metastases (Fig 1), in all three CCCs, in the CHCC-CC (Fig 2), in the undifferentiated carcinoma, in all four moderately or poorly differentiated HCCs (G2–G3, Fig 3), in two of the four highly differentiated HCCs (G1, Fig 4b), in the adenoma with atypia, and in five cysts. Three patients with G2–G3 HCCs and one of the two patients with G1 HCCs with absence of hepatocyte-selective enhancement had Child-Pugh class A cirrhosis.

View larger version (136K): [in this window] [in a new window] [Download PPT slide]   Figure 2a. CHCC-CC without hepatocyte-selective enhancement after gadoxetic acid injection in 66-year-old woman without cirrhosis. (a) Transverse T1-weighted GRE MR images acquired during arterial dominant phase. (b) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (c) Transverse T1-weighted GRE MR image acquired during hepatocyte-selective phase 20 minutes after injection of gadoxetic acid. During arterial dominant phase (a, 141.8/4.1; flip angle, 75°), lesion (arrow) in left liver lobe enhances and is nearly isointense to surrounding parenchyma. Central necrotic areas without enhancement can be seen. On CT scan (b), lesion (arrow) shows inhomogeneous enhancement and is nearly isoattenuating to surrounding parenchyma. Central hypoattenuating necrotic areas can be delineated, similar to MR imaging. In hepatocyte-selective phase 20 minutes after injection (c), well-delineated lesion without enhancement can be seen with T1-weighted fat-suppressed sequence (124.4/4.8; flip angle, 75°). (d) Photomicrograph of resected specimen demonstrates cholangiocarcinoma area of the CHCC-CC () with atypical ductular structures invading (arrows) normal liver parenchyma (#). Hepatocellular areas of CHCC-CC are not shown. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (141K): [in this window] [in a new window] [Download PPT slide]   Figure 2b. CHCC-CC without hepatocyte-selective enhancement after gadoxetic acid injection in 66-year-old woman without cirrhosis. (a) Transverse T1-weighted GRE MR images acquired during arterial dominant phase. (b) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (c) Transverse T1-weighted GRE MR image acquired during hepatocyte-selective phase 20 minutes after injection of gadoxetic acid. During arterial dominant phase (a, 141.8/4.1; flip angle, 75°), lesion (arrow) in left liver lobe enhances and is nearly isointense to surrounding parenchyma. Central necrotic areas without enhancement can be seen. On CT scan (b), lesion (arrow) shows inhomogeneous enhancement and is nearly isoattenuating to surrounding parenchyma. Central hypoattenuating necrotic areas can be delineated, similar to MR imaging. In hepatocyte-selective phase 20 minutes after injection (c), well-delineated lesion without enhancement can be seen with T1-weighted fat-suppressed sequence (124.4/4.8; flip angle, 75°). (d) Photomicrograph of resected specimen demonstrates cholangiocarcinoma area of the CHCC-CC () with atypical ductular structures invading (arrows) normal liver parenchyma (#). Hepatocellular areas of CHCC-CC are not shown. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (142K): [in this window] [in a new window] [Download PPT slide]   Figure 2c. CHCC-CC without hepatocyte-selective enhancement after gadoxetic acid injection in 66-year-old woman without cirrhosis. (a) Transverse T1-weighted GRE MR images acquired during arterial dominant phase. (b) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (c) Transverse T1-weighted GRE MR image acquired during hepatocyte-selective phase 20 minutes after injection of gadoxetic acid. During arterial dominant phase (a, 141.8/4.1; flip angle, 75°), lesion (arrow) in left liver lobe enhances and is nearly isointense to surrounding parenchyma. Central necrotic areas without enhancement can be seen. On CT scan (b), lesion (arrow) shows inhomogeneous enhancement and is nearly isoattenuating to surrounding parenchyma. Central hypoattenuating necrotic areas can be delineated, similar to MR imaging. In hepatocyte-selective phase 20 minutes after injection (c), well-delineated lesion without enhancement can be seen with T1-weighted fat-suppressed sequence (124.4/4.8; flip angle, 75°). (d) Photomicrograph of resected specimen demonstrates cholangiocarcinoma area of the CHCC-CC () with atypical ductular structures invading (arrows) normal liver parenchyma (#). Hepatocellular areas of CHCC-CC are not shown. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (196K): [in this window] [in a new window] [Download PPT slide]   Figure 2d. CHCC-CC without hepatocyte-selective enhancement after gadoxetic acid injection in 66-year-old woman without cirrhosis. (a) Transverse T1-weighted GRE MR images acquired during arterial dominant phase. (b) Corresponding transverse multi-detector row CT scan obtained in arterial phase. (c) Transverse T1-weighted GRE MR image acquired during hepatocyte-selective phase 20 minutes after injection of gadoxetic acid. During arterial dominant phase (a, 141.8/4.1; flip angle, 75°), lesion (arrow) in left liver lobe enhances and is nearly isointense to surrounding parenchyma. Central necrotic areas without enhancement can be seen. On CT scan (b), lesion (arrow) shows inhomogeneous enhancement and is nearly isoattenuating to surrounding parenchyma. Central hypoattenuating necrotic areas can be delineated, similar to MR imaging. In hepatocyte-selective phase 20 minutes after injection (c), well-delineated lesion without enhancement can be seen with T1-weighted fat-suppressed sequence (124.4/4.8; flip angle, 75°). (d) Photomicrograph of resected specimen demonstrates cholangiocarcinoma area of the CHCC-CC () with atypical ductular structures invading (arrows) normal liver parenchyma (#). Hepatocellular areas of CHCC-CC are not shown. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (179K): [in this window] [in a new window] [Download PPT slide]   Figure 3a. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (165K): [in this window] [in a new window] [Download PPT slide]   Figure 3b. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (151K): [in this window] [in a new window] [Download PPT slide]   Figure 3c. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (162K): [in this window] [in a new window] [Download PPT slide]   Figure 3d. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (190K): [in this window] [in a new window] [Download PPT slide]   Figure 3e. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (174K): [in this window] [in a new window] [Download PPT slide]   Figure 3f. G2-G3 HCC without hepatocyte-selective enhancement after gadoxetic acid injection and G1 HCC with enhancement in 73-year-old man with Child-Pugh class A cirrhosis and two HCCs. Transverse T1-weighted fat-suppressed GRE MR images (124.4/4.8; flip angle, 75°) (a, b) before and (c, d) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Before injection, isointense lesion in left liver lobe (arrow) can be seen (a). On anatomically inferior section (b), nodular transformation of liver parenchyma can be observed (arrowheads), but no focal lesions can be clearly delineated. In hepatocyte-selective phase (c), no relevant enhancement of lesion in left liver lobe can be detected, and lesion is surrounded by capsule with homogeneous enhancement (arrowheads). On inferior section (d), well-delineated lesion not visible on precontrast image can be seen (arrow). Contrary to lesion in left liver lobe, this lesion is hyperintense in comparison to surrounding parenchyma. (e, f) Photomicrographs of resected specimens confirmed suspicion of different levels of differentiation in both lesions. Histopathologic image of large lesion (e) revealed moderately to poorly differentiated HCC (G2-G3), whereas small lesion (f) was graded as highly differentiated in part with clear cell aspect. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (160K): [in this window] [in a new window] [Download PPT slide]   Figure 4a. G1 HCC without hepatocyte-selective enhancement after gadoxetic acid injection in 69-year-old man without cirrhosis with HCC in right liver lobe. Transverse T1-weighted GRE MR images acquired with fat suppression (124.4/4.8; flip angle, 75°) at 1.5 T (a) 10 and (b) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Note large hypointense and inhomogeneous mass (arrow) with large necrotic areas. No hepatocyte-selective enhancement is visible in HCC 10 and 20 minutes after injection. Note biliary enhancement in common hepatic duct 20 minutes after injection (arrow) not present on image acquired 10 minutes after injection. (c) Histopathologic examination demonstrated G1 HCC with predominantly pseudoglandular pattern (arrows) and nearly bland nuclei. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (178K): [in this window] [in a new window] [Download PPT slide]   Figure 4b. G1 HCC without hepatocyte-selective enhancement after gadoxetic acid injection in 69-year-old man without cirrhosis with HCC in right liver lobe. Transverse T1-weighted GRE MR images acquired with fat suppression (124.4/4.8; flip angle, 75°) at 1.5 T (a) 10 and (b) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Note large hypointense and inhomogeneous mass (arrow) with large necrotic areas. No hepatocyte-selective enhancement is visible in HCC 10 and 20 minutes after injection. Note biliary enhancement in common hepatic duct 20 minutes after injection (arrow) not present on image acquired 10 minutes after injection. (c) Histopathologic examination demonstrated G1 HCC with predominantly pseudoglandular pattern (arrows) and nearly bland nuclei. (Hematoxylin-eosin stain; original magnification, x200.)

View larger version (203K): [in this window] [in a new window] [Download PPT slide]   Figure 4c. G1 HCC without hepatocyte-selective enhancement after gadoxetic acid injection in 69-year-old man without cirrhosis with HCC in right liver lobe. Transverse T1-weighted GRE MR images acquired with fat suppression (124.4/4.8; flip angle, 75°) at 1.5 T (a) 10 and (b) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Note large hypointense and inhomogeneous mass (arrow) with large necrotic areas. No hepatocyte-selective enhancement is visible in HCC 10 and 20 minutes after injection. Note biliary enhancement in common hepatic duct 20 minutes after injection (arrow) not present on image acquired 10 minutes after injection. (c) Histopathologic examination demonstrated G1 HCC with predominantly pseudoglandular pattern (arrows) and nearly bland nuclei. (Hematoxylin-eosin stain; original magnification, x200.)

On the other hand, liver-specific enhancement was seen in the two other G1 HCCs, in two adenomas, in all three cases of FNH, and in the biliary cystadenoma. Both G1 HCCs with enhancement were seen in patients with underlying Child-Pugh class A or B cirrhosis. All cases of FNH exhibited heterogeneous enhancement in the liver-specific phases and were hyperintense in comparison to the normal liver parenchyma. In the two large FNHs (maximum diameter > 10 cm), nonenhancing central regions and septae were demonstrated (Fig 5). At histopathologic examination, these areas corresponded to a fibrous scar. The third FNH was smaller (maximum diameter of 3 cm) and showed heterogeneous hyperintensity with central isointense patches at MR imaging (Fig 5) and no prominent central scar at histopathologic examination. Two of the three adenomas were hyperintense: one with homogeneous enhancement, the other with a heterogeneous hyperintensity with central isointense spots. In the two G1 HCCs with hepatocyte-selective uptake, the first was isointense with a homogeneous enhancement, and the second was centrally hyperintense with a small peripheral hypointense rim (Fig 3).

View larger version (125K): [in this window] [in a new window] [Download PPT slide]   Figure 5a. Two FNHs with hepatocyte-selective enhancement after gadoxetic acid injection in 38-year-old woman with chronic pain in upper abdomen. Hemihepatectomy of left liver lobe for resection of large lesion (diameter, 11 cm) was performed. For smaller lesion in right liver lobe, core-needle biopsy was performed with same histopathologic finding. Transverse MR images acquired at 1.5 T (a) before contrast material injection, (b) during arterial dominant phase, and (c) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Precontrast T1-weighted GRE MR image with fat suppression (124.4/4.8; flip angle, 75°) (a) shows isointense large lesion (arrow) invading entire left liver lobe. Arterial dominant phase image (b) (T1-weighted GRE image without fat suppression; 141.8/4.1; flip angle, 75°) depicted second lesion in right liver lobe (arrow) and two homogeneously hyperintense lesions. In hepatocyte-specific phase (c) 20 minutes after injection (same sequence as a), both lesions are slightly hyperintense in comparison to normal liver parenchyma. Lesion in left liver lobe shows central hypointensity (arrow) corresponding to central scar radiating toward periphery. Lesion in right liver lobe (arrowhead) shows heterogeneous hyperintensity with central isointense patches. (d) Photomicrograph of resected specimen demonstrates characteristic findings of FNH with central fibrosis (), chronic inflammation, and nodular hepatocellular (arrowhead) and neoductular (arrows) proliferation. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (126K): [in this window] [in a new window] [Download PPT slide]   Figure 5b. Two FNHs with hepatocyte-selective enhancement after gadoxetic acid injection in 38-year-old woman with chronic pain in upper abdomen. Hemihepatectomy of left liver lobe for resection of large lesion (diameter, 11 cm) was performed. For smaller lesion in right liver lobe, core-needle biopsy was performed with same histopathologic finding. Transverse MR images acquired at 1.5 T (a) before contrast material injection, (b) during arterial dominant phase, and (c) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Precontrast T1-weighted GRE MR image with fat suppression (124.4/4.8; flip angle, 75°) (a) shows isointense large lesion (arrow) invading entire left liver lobe. Arterial dominant phase image (b) (T1-weighted GRE image without fat suppression; 141.8/4.1; flip angle, 75°) depicted second lesion in right liver lobe (arrow) and two homogeneously hyperintense lesions. In hepatocyte-specific phase (c) 20 minutes after injection (same sequence as a), both lesions are slightly hyperintense in comparison to normal liver parenchyma. Lesion in left liver lobe shows central hypointensity (arrow) corresponding to central scar radiating toward periphery. Lesion in right liver lobe (arrowhead) shows heterogeneous hyperintensity with central isointense patches. (d) Photomicrograph of resected specimen demonstrates characteristic findings of FNH with central fibrosis (), chronic inflammation, and nodular hepatocellular (arrowhead) and neoductular (arrows) proliferation. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (122K): [in this window] [in a new window] [Download PPT slide]   Figure 5c. Two FNHs with hepatocyte-selective enhancement after gadoxetic acid injection in 38-year-old woman with chronic pain in upper abdomen. Hemihepatectomy of left liver lobe for resection of large lesion (diameter, 11 cm) was performed. For smaller lesion in right liver lobe, core-needle biopsy was performed with same histopathologic finding. Transverse MR images acquired at 1.5 T (a) before contrast material injection, (b) during arterial dominant phase, and (c) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Precontrast T1-weighted GRE MR image with fat suppression (124.4/4.8; flip angle, 75°) (a) shows isointense large lesion (arrow) invading entire left liver lobe. Arterial dominant phase image (b) (T1-weighted GRE image without fat suppression; 141.8/4.1; flip angle, 75°) depicted second lesion in right liver lobe (arrow) and two homogeneously hyperintense lesions. In hepatocyte-specific phase (c) 20 minutes after injection (same sequence as a), both lesions are slightly hyperintense in comparison to normal liver parenchyma. Lesion in left liver lobe shows central hypointensity (arrow) corresponding to central scar radiating toward periphery. Lesion in right liver lobe (arrowhead) shows heterogeneous hyperintensity with central isointense patches. (d) Photomicrograph of resected specimen demonstrates characteristic findings of FNH with central fibrosis (), chronic inflammation, and nodular hepatocellular (arrowhead) and neoductular (arrows) proliferation. (Hematoxylin-eosin stain; original magnification, x100.)

View larger version (177K): [in this window] [in a new window] [Download PPT slide]   Figure 5d. Two FNHs with hepatocyte-selective enhancement after gadoxetic acid injection in 38-year-old woman with chronic pain in upper abdomen. Hemihepatectomy of left liver lobe for resection of large lesion (diameter, 11 cm) was performed. For smaller lesion in right liver lobe, core-needle biopsy was performed with same histopathologic finding. Transverse MR images acquired at 1.5 T (a) before contrast material injection, (b) during arterial dominant phase, and (c) 20 minutes after administration of 0.025 mmol/kg gadoxetic acid. Precontrast T1-weighted GRE MR image with fat suppression (124.4/4.8; flip angle, 75°) (a) shows isointense large lesion (arrow) invading entire left liver lobe. Arterial dominant phase image (b) (T1-weighted GRE image without fat suppression; 141.8/4.1; flip angle, 75°) depicted second lesion in right liver lobe (arrow) and two homogeneously hyperintense lesions. In hepatocyte-specific phase (c) 20 minutes after injection (same sequence as a), both lesions are slightly hyperintense in comparison to normal liver parenchyma. Lesion in left liver lobe shows central hypointensity (arrow) corresponding to central scar radiating toward periphery. Lesion in right liver lobe (arrowhead) shows heterogeneous hyperintensity with central isointense patches. (d) Photomicrograph of resected specimen demonstrates characteristic findings of FNH with central fibrosis (), chronic inflammation, and nodular hepatocellular (arrowhead) and neoductular (arrows) proliferation. (Hematoxylin-eosin stain; original magnification, x100.)

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
In clinical studies, liver-specific contrast media have been demonstrated to improve markedly the detection of focal liver lesions (1,6). In special indications, such as preoperative evaluation of patients with HCCs or metastases, the superiority of these contrast media has also been proved with regard to patient management and cost-effectiveness in comparison to biphasic spiral CT (18). In contrast to the common scientific approach of evaluating lesion detection and characterization separately, the subsequent characterization of a newly detected focal liver lesion is of central importance in daily medical routine.

To evaluate the performance of a bolus-injectable liver-specific agent for characterization, perfusion properties and liver-specific properties should be evaluated separately to assess the additional information provided by the liver-specific phase. Our results demonstrate a statistically significant high correlation of radiomorphologic and enhancement features in the arterial and portal venous phases between gadoxetic acid–enhanced MR imaging and CT when using a state-of-the-art examination technique. Thus, evidence was provided of at least equal results in the detection and characterization of lesions during arterial and portal venous phases and in the confirmation of conclusions of previous publications on the comparison of differential diagnosis information of gadoxetic acid–enhanced MR imaging with examinations in which an unspecific extracellular contrast medium was used (7,8). This is particularly important for gadoxetic acid, since the injected dose of gadolinium (0.025 mmol/kg) is lower than the usual dose of extracellular gadolinium chelates (0.1 mmol/kg), although partially compensated for by a higher relaxivity (3).

As for the hepatocyte-selective phases, uptake of hepatocyte-selective agents occurs not only in normal liver parenchyma but also in focal liver lesions (adenoma, FNH, HCC) of hepatocellular origin (7,19–23). Thus, the possibility arises of distinguishing hepatocyte-containing from non–hepatocyte-containing tissue. A previous publication (19) indicated the superiority of mangafodipir trisodium to spiral CT in this differential diagnosis. Nevertheless, the potential of mangafodipir trisodium is limited by the possible "paradoxic uptake" in metastases (24).

Gadoxetic acid promises to be more effective, since a paradoxic uptake in metastases was not observed in our collective, and, to the best of our knowledge, was never reported in the literature. Our results suggest that the presence of uptake in the hepatocyte-selective phases strongly indicates a lesion originating from hepatocytes. Thus, all benign or tumorlike hepatocellular lesions in our collective (FNH, two adenomas, and the biliary cystadenoma), with the exception of the adenoma with atypia, displayed hepatocyte-selective uptake. Thus, results confirmed that these lesion types enhance regularly after injection of gadobenate dimeglumine (21,22) and mangafodipir trisodium (23). Furthermore, the focal lack of enhancement in some areas in FNH correlating with histopathologically proved scars confirms that no enhancement is to be expected in areas without hepatocytes.

As for malignant liver tumors, HCCs are suggested to derive from hepatocytes, and CCCs from ductular epithelia (25). The origin of CHCC-CC, however, as well as that of undifferentiated carcinoma of the liver, remains to be clarified. These entities and the sometimes striking intratumorous heterogeneity gave rise to the still unproven theory of a putative common progenitor cell that generates a more hepatocellular or more cholangiocellular differentiation within the same tumor.

Our results confirm the early preclinical findings that the hepatocyte-selective uptake of gadoxetic acid reflects the cellular origin of tumors and might therefore be limited to highly differentiated malignant hepatocellular lesions (26,27). No relevant uptake in CCC, undifferentiated carcinoma, or CHCC-CC was observed. In HCC, the expression of the organic anion-transporting polypeptide 1 carrier could be correlated with the degree of differentiation. In a publication without special focus on tumor grading (28), a genetic underexpression of key gene products, including organic anion-transporting polypeptide, in surgically resected human HCC in comparison to the surrounding tissue was reported. The underexpression was considered to be important in the development and/or progression of HCC. Various authors have based their approach to grading HCC in vivo in a quantitative MR evaluation of hepatocyte-specific contrast material uptake on this assumption (26,27,29,30).

For gadoxetic acid, the mechanism of uptake is similar to that known for the hepatobiliary radioactive tracer technetium 99m (^99mTc)–labeled iminodiacetic acid, or IDA (31,32). In a previous publication, Van Beers et al (33) compared the uptake of gadoxetic acid and the hepatobiliary radioactive tracer ^99mTc IDA in chemically induced HCC in rats. They found an exceeding uptake with IDA in these tumors in comparison to the normal liver parenchyma but a low uptake of gadoxetic acid and concluded that the enhancement pattern of well to moderately differentiated HCCs (classified according to Squire and Levitt [34] and the Institute of Laboratory Animal Resources [35]) with gadoxetic acid does not mirror that observed with ^99mTc IDA.

In humans, in previous publications involving 12 (7) and four (8) patients with histopathologically proved HCCs, isointense or even hyperintense HCCs were not observed in the hepatocyte-selective phases after injection of gadoxetic acid at MR imaging (7,8). However, the degree of differentiation of the tumors was not specified. In our collective, two G1 HCCs in patients with liver cirrhosis exhibited an exceeding or equal uptake in comparison to the surrounding parenchyma. We therefore cannot confirm reports (7,8,29,33) that HCCs, independent from their differentiation, enhance less than normal liver after gadoxetic acid.

Our results are supported by two previous publications (26,27) based on animal models with chemically induced HCCs. Both publications involved the use of the classification system of Edmondson and Steiner (11). In one publication (26), the two well-differentiated HCCs were hyperintense compared with the surrounding parenchyma, whereas all of the 77 moderately differentiated or undifferentiated HCCs remained unenhanced. In the other study (27), pronounced positive enhancement was seen in eight of 66 hepatomas. All eight lesions were classified as well differentiated.

In another animal study (37) with 66 HCCs, however, positive enhancement in two well differentiated and two moderately to poorly differentiated HCCs (classified according to Vesselinivitch [36], practically corresponding to grades G1–G3 according to World Health Organization criteria) was reported. Our study results reveal no uptake in moderately or poorly differentiated HCC in humans.

Generally, it had to be taken into consideration that animal results in induced in vivo neoplasm cannot be transferred directly to human HCC, since the influences of induced carcinogenesis on the molecular environment—that is, carriers and enzymes—are not known. Therefore, while absent in hepatoma cell lines, organic anion-transporting polypeptide 1 could still be expressed in well-differentiated HCC in vivo. Our results might suggest that relevant uptake in a lesion does not exclude malignancy, but, if the lesion is malignant, it might favor an HCC with a higher degree of differentiation. The absence of enhancement in our adenoma with atypia, which to the best of our knowledge has not been described in previous publications involving the use of hepatocyte-selective MR imaging contrast media, underlines the variability of cellular differentiation. In the hepatocyte-selective phases of gadoxetic acid–enhanced MR imaging, there is no clearly defined point of intersection between benign and malignant lesions.

Our results underline that the differential diagnosis cannot be assigned only in the hepatocyte-selective phases but that precontrast and dynamic MR sequences are indispensable for the diagnosis. In addition, detection of well-differentiated HCCs