(Radiology. 2001;218:27-38.)
© RSNA, 2001
Contrast Agents for MR Imaging of the Liver1
Richard C. Semelka, MD and
Thomas K. G. Helmberger, MD
1 From the Department of Radiology, University of North Carolina School of Medicine, CB 7510, Chapel Hill, NC 27599-7510 (R.C.S.); and the Department of Radiology, Klinikum Grosshadern, Munich, Germany (T.K.G.H.). Received June 22, 1999; revision requested August 6; final revision received November 15; accepted December 6, 1999. Address correspondence to R.C.S. (e-mail: richsem@med.unc.edu).
 |
ABSTRACT
|
|---|
A variety of different categories of contrast agents, and within each category a number of individual agents, are currently available for clinical use in magnetic resonance (MR) imaging of the liver. In this review, the use of nonspecific extracellular gadolinium chelates, reticuloendothelial system–specific iron oxide particulate agents, hepatocyte-selective agents, and combined perfusion and hepatocyte-selective agents are described. Most clinical experience is with nonspecific extracellular gadolinium chelates. The relatively low cost, safety, good patient tolerance, and ability to help detect and characterize a wide range of liver diseases have rendered gadolinium chelates as commonly used agents. Reticuloendothelial system–specific agents improve lesion detection by decreasing the signal intensity of background liver on T2-weighted MR images, which increases the conspicuity of focal hepatic lesions with negligible reticuloendothelial cells (eg, metastases). Hepatocyte-selective agents increase the signal intensity of background liver on T1-weighted images, which increases the conspicuity of focal lesions that do not contain hepatocytes (eg, metastases). The clinical application of the different categories of contrast agents, techniques for their administration, sequences to be used, and appearances of common entities on contrast agent–enhanced studies are described.
Index terms: Liver, MR, 761.1214 • Liver neoplasms, 761.3198, 761.32, 761.33 • Magnetic resonance (MR), contrast media, 761.12143 • State of the Art
|
INTRODUCTION
|
|---|
Intravenously administered contrast agents have been used in clinical magnetic resonance (MR) imaging of the liver since approximately 1988. The first category of contrast agents to be used in clinical practice was that of nonspecific extracellular gadolinium chelates. Since then, two other classes of contrast agents have been developed for liver MR studies: reticuloendothelial system (RES)–specific contrast agents and hepatocyte-selective contrast agents. In this article, we will describe agents in each of these three categories: nonspecific extracellular, RES-specific, and hepatocyte-selective contrast agents. Emphasis is placed on nonspecific extracellular gadolinium chelates because they have been available for the longest time and have the best documented safety profile and clinical applications.
In determining the role for contrast agents, an understanding of the information available with use of nonenhanced MR sequences is important to facilitate appreciation of what additional information is provided when contrast material is used.
|
NONENHANCED MR SEQUENCES
|
|---|
T1-weighted Sequences
Breath-hold spoiled gradient-echo (GRE) imaging has achieved widespread use as a T1-weighted sequence for evaluation of the liver (1). Typical parameters include a repetition time (TR) of 150 msec, an echo time (TE) of 4 msec (at 1.5 T), a flip angle of 80°, a section thickness of 8 mm, an intersection gap of 20%, a matrix of 128–171 x 256 (phase-encoding by frequency-encoding), and an acquisition time of 14–22 seconds with a 20-second breath hold. The TR is sufficiently long to acquire enough sections to cover the entire liver in one pass and to provide a good signal-to-noise ratio. The TE is the shortest in-phase TE, which provides strong T1 weighting, minimizes magnetic susceptibility effects, and permits acquisition of a sufficient number of sections to cover the entire liver. A flip angle of 80° provides good T1 weighting and less radio-frequency power deposition and tissue saturation than would a larger flip angle that would provide comparable T1 weighting. Matrix size and section thickness are adequate to provide good in-plane and through-plane resolution, respectively.
The parameters for spoiled GRE imaging appear to be generalizable between MR machines produced by different manufacturers, such that the images acquired with sequences performed by using the above parameters with any equipment that has gradient capability would be of good diagnostic quality. An important feature of the multisection acquisition of spoiled GRE imaging is that the central phase-encoding steps (which determine the bulk signal for the image) are acquired during 6 seconds for both the entire data set and each individual section. The data acquisition time is, therefore, sufficiently short for the entire data set to isolate a distinct phase of gadolinium enhancement (eg, hepatic arterial–dominant phase); at the same time, the data acquisition time of each individual section is sufficiently long to compensate for slight variations in patient cardiac output, peak lesion enhancement, and injection technique.
Currently, three-dimensional GRE MR imaging is being investigated as a T1-weighted sequence for the liver (2). The benefits of this technique include an increased signal-to-noise ratio, as compared with that for two-dimensional techniques, and three-dimensional volume acquisition, which permits high-spatial-resolution multiplanar display. The technique is particularly useful for defining the hepatic blood vessels (2). The image quality obtained with three-dimensional GRE techniques is not yet equivalent to that obtained with two-dimensional spoiled GRE methods, and further refinements are ongoing.
Magnetization-prepared GRE sequences are used to acquire data with a single-section technique, with each individual section acquired in less than 2 seconds (1). This sequence is relatively insensitive to artifacts from patient motion and breathing, and it has thus been termed breathing independent. T1-weighted spin-echo studies are generally acquired over 2–6 minutes and are therefore obtained while the patient is breathing; thus, this sequence is termed breathing averaged.
T1-weighted images demonstrate imaging features for various types of liver lesions. Lesions with a high fluid content (eg, cyst, hemangioma) have very low signal intensity, lesions that are hypovascular or have a high fibrous tissue content (eg, colon cancer metastases, transitional cell carcinoma metastases, antibiotic-treated hepatic infection lesions, chemotherapy-treated metastases that have undergone changes of fibrosis, hepatic fibrosis) have moderately low signal intensity. Lesions that are hemorrhagic (eg, hemorrhagic metastases, liver hemorrhage), have high protein content (eg, hepatocellular carcinoma), have high fat content (eg, hepatocellular carcinoma, hepatic adenoma), or contain melanin (eg, melanoma) have high signal intensity (3,4). This information provided by nonenhanced T1-weighted imaging aids in lesion detection and characterization.
T2-weighted Sequences
Breathing-averaged T2-weighted conventional or echo-train (eg, fast or turbo) spin-echo MR sequences, often combined with fat suppression, are the most commonly used T2-weighted sequences. In patients who breathe in a regular fashion, these sequences generate images of high diagnostic quality. Many patients, however, are not able to breathe regularly, and the image quality of breathing-averaged sequences is inconsistent. Therefore, breath-hold or breathing-independent T2-weighted MR imaging is now being used at many centers in place of breathing-averaged T2-weighted imaging. Unlike T1-weighted spoiled GRE sequences, the optimized imaging parameters for T2-weighted breath-hold or breathing-independent sequences have not been realized as yet, particularly among equipment manufacturers, and we will not describe recommended parameters.
T2-weighted images provide information on fluid content (reflected by high signal intensity) and iron content (reflected by low signal intensity) (3). It has been recognized that high fluid content in focal liver lesions has a strong correlation with benign disease (ie, cysts and hemangiomas), whereas lower fluid content is more typical of malignant disease (5). Many investigators have described the value of distinguishing cysts and hemangiomas from metastases and hepatocellular carcinoma on the basis of their appearance on T2-weighted images. Sole reliance on T2-weighted images may not be advisable, however, because some liver metastases that are either cystic (eg, ovarian cancer metastases) or hypervascular (eg, gastrointestinal leiomyosarcoma metastases, islet cell tumor metastases) have a high fluid content and therefore have high signal intensity on T2-weighted images (5–7). On the other hand, relatively common benign liver lesions such as focal nodular hyperplasia and hepatic adenoma have a relatively low fluid content and are therefore approximately isointense relative to liver on T2-weighted images.
Sequence Modifications
Excitation-spoiling chemically selective fat suppression is effective at diminishing phase artifacts due to respiration and improves the dynamic range of signal intensities of abdominal tissues. Fat suppression is particularly effective at improving the visualization of contrast enhancement in regions bordered by fat (eg, subcapsular liver). Echo-train T2-weighted sequences also benefit from the use of fat suppression, especially in the setting of fatty liver. In-phase and out-of-phase T1-weighted spoiled GRE sequences are routinely used in liver studies to evaluate for fatty liver.
Coil Use
Phased-array or multiarray torso coils are a useful addition to MR imaging of the liver. The improvement in signal-to-noise ratio permits the acquisition of thinner sections and higher diagnostic quality of many short-duration MR sequences. The benefit of their implementation is greatest in patients with a lean to medium build. Substantial signal loss in the central portion of the abdomen may detract from image quality either when the patient is large or when adequate signal normalization postprocessing has not been performed.
|
CONTRAST-ENHANCED MR SEQUENCES
|
|---|
Why Contrast Agents?
The need for more accurate characterization of all histologic types of liver lesions and for detection of the full extent of malignant liver lesions have been major reasons for the use of intravenous contrast agents (8,9). The additional need to detect and characterize extrahepatic disease has developed into an equally important reason to use contrast agents (10). The requirement for contrast material has apparently not diminished despite the improved image quality possible with newer nonenhanced T1- and T2-weighted sequences for investigating the liver.
Nonspecific Extracellular Gadolinium Chelates
The use of these agents has been considered essential for evaluation of the full complexity of abdominal disease in patients evaluated with MR imaging for a diverse range of indications. To facilitate the broad application of abdominal MR imaging, the use of standardized imaging protocols may be advisable (11). A combination of various sequences that include breath-hold or breathing-independent sequences shortens total examination time. Comprehensive diagnostic information is provided, encompassing the full range of abdominal diseases, by using a variety of types of sequences, different planes, and the routine use of gadolinium chelates. The sequence list should be sufficiently short to avoid redundancy and excessive examination time. Table 1 illustrates a standard protocol for MR imaging of the liver. In our clinical experience, it is difficult to predict prospectively who can be adequately imaged with a nonenhanced MR examination of the liver. As a result, our practice is to use gadolinium-enhanced MR imaging as the standard routine MR examination of the liver. The most compelling case for the routine use of gadolinium chelates is examination of the liver in patients who are suspected of having hepatocellular carcinoma or hypervascular malignant tumors (12,13).
Gadolinium chelates are optimally used when administered as a rapid bolus, and imaging is performed with a T1-weighted spoiled GRE sequence that is repeated in a dynamic serial fashion (5–9,12–16). This is best achieved at high field strength (1.0 or 1.5 T). Our practice is to use the shortest in-phase sequence. It is also acceptable to use a shorter TE sequence with an out-of-phase TE. It may be prudent to not use an out-of-phase sequence in patients with a fatty-infiltrated liver, because signal intensity variation in the liver may confound the interpretation of enhancement. The other problem with an out-of-phase sequence is that a phase-cancellation artifact will mask the liver capsule, which may render the detection of capsular or small subcapsular lesions difficult. This can be avoided by the concomitant use of fat suppression. A minimum of two nonenhanced acquisitions may be essential, and little additional information is provided if more than four are acquired. Our standard is to perform three acquisitions by using the following timing of phases.
Hepatic arterial–dominant phase.—The hepatic arterial–dominant phase, also termed the capillary phase when other organs are imaged, is the single most important MR data set when a nonspecific extracellular gadolinium chelate is used (12,13). This phase of enhancement is the only postcontrast sequence for which timing is crucial. It is essential to capture the "first pass," or capillary bed enhancement, of tissues during this phase. Demonstration of gadolinium enhancement in hepatic arteries and portal veins and absence of gadolinium enhancement in hepatic veins are reliable landmarks (Fig 1). In general, if contrast material is injected at a rate of 2–3 mL/sec by means of a power injector or by hand, this phase is achieved in the majority of patients by initiating a spoiled GRE sequence with standard ordering of the phase-encoding table at 16–17 seconds after the start of injection. Alternatively, if rapid-bolus hand injection is performed by injecting 0.1 mmol of gadolinium per kilogram body weight (typically a 15–20-mL dose) followed by rapid injection of a 10-mL bolus of normal saline and initiation of a spoiled GRE sequence after completion of the normal saline bolus, consistent results are also obtained. Reproducibility may be further improved by using a test bolus to calculate more exact timing (17).
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Transverse spoiled GRE MR image (150/4, 80° flip angle) demonstrates normal hepatic arterial-dominant phase of gadolinium enhancement. High-signal-intensity gadolinium enhancement is present in portal veins (large straight arrow) and hepatic artery (short straight arrow) but not in hepatic veins (curved arrow).
|
|
In the hepatic arterial–dominant phase of enhancement, we believe that, although gadolinium enhancement is present in portal veins, the majority of the contrast material in the liver has been delivered by hepatic arteries (3). The absolute volume of contrast material delivered via the hepatic arteries is greater during this phase of enhancement than if the data are acquired when contrast material is present in the hepatic arteries alone, which means that more hepatic arterial blood supply information is available. This is important since most focal lesions, especially metastases and hepatocellular carcinoma, are fed primarily by hepatic arteries (7).
Imaging slightly earlier than this phase, when only hepatic arteries are opacified (hepatic arteries–only phase), may approach the diagnostic utility of the hepatic arterial–dominant phase if the injection rate of contrast material is fast and the sequence is acquired late in the hepatic arteries–only phase (within a very short time of contrast material arrival in the portal veins). Observation of high signal intensity in the renal cortex (approaching that of the aorta), pancreas, and spleen reflects adequate contrast material delivery during the hepatic arterial–only phase of enhancement. It is difficult to achieve these objectives in the hepatic arterial–only phase, and it is also difficult to judge if image acquisition is too early in this phase, at a time when the liver is essentially not enhanced.
Appropriate timing, as judged by observing vessel enhancement, also is important for the evaluation of surrounding organs. Negligible or minimal pancreatic enhancement (ie, minimal increase in signal intensity, as compared with nonenhanced images) is consistently observed in pancreatic fibrosis or chronic pancreatitis, and negligible or minimal enhancement of the renal cortex may imply ischemic nephropathy or acute cortical necrosis. This can be reliably judged on hepatic arterial–dominant phase images, due to the fixed landmarks of contrast enhancement in portal veins and the absence of enhancement in hepatic veins. In the hepatic arterial–only phase, minimal enhancement of the pancreas or renal cortex may reflect too-early image acquisition rather than a disease process. Since this immediate postenhancement phase is also used to help diagnose adequate perfusion of these organs, it may be problematic to use enhancement of these organs as the means of determining the appropriateness of the phase of image acquisition timing. Enhancement of the pancreas or renal cortex provides useful ancillary information for the appropriateness of timing, although it is not the major determinant, because the extent of enhancement of these organs is also evaluated at this phase. In the liver, imaging too early in the hepatic arterial–only phase diminishes the ability to recognize the distinctive patterns of lesion enhancement for different lesion types, because the absolute volume of contrast material delivered via the hepatic artery may be too small and may cause lesions to be mischaracterized on the basis of minimal lesion enhancement (Fig 2).
View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Transverse gadolinium-enhanced spoiled GRE MR images (150/4, 180° flip angle) obtained during the (a) hepatic arterial-only and (b) hepatic arterial-dominant phases in a patient with two focal nodular hyperplasia lesions. (a) Minimal heterogeneous enhancement is apparent in these lesions, which thereby raised the suspicion that the lesions may represent metastases or hepatocellular carcinoma. (b) Subsequently, a correctly timed hepatic arterial-dominant phase image demonstrates that the lesions possess uniform intense enhancement, with lack of enhancement of the small central scars (arrows). This appearance in conjunction with the T2-weighted imaging (not shown) appearance of mild hyperintensity of the lesions and moderate hyperintensity of the central scars is virtually diagnostic of focal nodular hyperplasia.
|
|
View larger version (140K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Transverse gadolinium-enhanced spoiled GRE MR images (150/4, 180° flip angle) obtained during the (a) hepatic arterial-only and (b) hepatic arterial-dominant phases in a patient with two focal nodular hyperplasia lesions. (a) Minimal heterogeneous enhancement is apparent in these lesions, which thereby raised the suspicion that the lesions may represent metastases or hepatocellular carcinoma. (b) Subsequently, a correctly timed hepatic arterial-dominant phase image demonstrates that the lesions possess uniform intense enhancement, with lack of enhancement of the small central scars (arrows). This appearance in conjunction with the T2-weighted imaging (not shown) appearance of mild hyperintensity of the lesions and moderate hyperintensity of the central scars is virtually diagnostic of focal nodular hyperplasia.
|
|
On hepatic arterial–dominant phase T1-weighted spoiled GRE images, various types of liver lesions have distinctive enhancement patterns: Cysts show lack of enhancement, hemangiomas show peripheral nodules of enhancement in a discontinuous ring, nonhemorrhagic adenomas and focal nodular hyperplasia show intense uniform enhancement, metastases show ring enhancement, and hepatocellular carcinomas show diffuse heterogeneous enhancement (Fig 3) (3,9, 15,16,18–20). The ability to use this information to characterize lesions as small as 1 cm in diameter may be most consistently achieved with MR imaging (Fig 4). Appearances of less common liver lesions on images obtained immediately after enhancement have also been reported, and many of these descriptions overlap with the above-described patterns of common liver lesions (21,22).
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3c. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3d. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3e. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (113K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3f. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show typical enhancement of various focal hepatic lesions. (a) A hepatic cyst has well-defined margins and is a signal void (arrow). (b) Hemangioma manifests as peripheral nodules (arrows) in a discontinuous ring. (c) Adenoma (arrow) shows intense uniform enhancement. (d) Focal nodular hyperplasia shows intense uniform enhancement with lack of enhancement of the central scar (arrow). (e) Metastases show ring enhancement (arrow). (f) Hepatocellular carcinoma (arrow) shows diffuse heterogeneous enhancement.
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show characteristic enhancement features of small (<1-cm) hepatic lesions. (a) Hemangioma shows peripheral nodules in a discontinuous ring (arrow). (b) Metastasis (arrow) shows peripheral ring enhancement.
|
|
View larger version (173K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Transverse gadolinium-enhanced spoiled GRE MR images (140-160/4, 80° flip angle) obtained during the hepatic arterial-dominant phase show characteristic enhancement features of small (<1-cm) hepatic lesions. (a) Hemangioma shows peripheral nodules in a discontinuous ring (arrow). (b) Metastasis (arrow) shows peripheral ring enhancement.
|
|
Knowledge of clinical history is often important despite the high diagnostic accuracy possible with current MR imaging protocols (18,22). In addition, many histologic types of lesions, when they measure less than 1 cm, demonstrate virtually identical uniform enhancement (eg, hemangiomas, adenomas, focal nodular hyperplasia, metastases, and hepatocellular carcinoma) (7,11,12,23). Ancillary information to assist in characterization of lesions is crucial. This information includes T2-weighted MR images that demonstrate lesion fluid content (eg, high signal intensity for cysts, hamartoma, and hemangioma and often high signal intensity for hypervascular metastases; relatively low signal intensity for hypovascular metastases, adenoma, focal nodular hyperplasia, and hepatocellular carcinoma) (Fig 5), other concomitant large lesions, and clinical history (eg, history of known primary tumor that can result in hypervascular metastases, including gastrointestinal leiomyosarcoma, or of underlying cirrhosis or hepatitis that would predispose to hepatocellular carcinoma) (7,18,23). Specific appearances of various types of metastases are also appreciable on hepatic arterial–dominant phase images. Colon cancer and pancreatic ductal adenocarcinoma metastases typically have ill-defined perilesional enhancement. In addition, colon cancer metastases larger than 3 cm in diameter commonly have a cauliflower-like appearance.
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Coronal T2-weighted single-shot echo-train spin-echo MR images ( /90) demonstrate fluid content of focal hepatic lesions. (a) In a patient with hemangioma, the high fluid content of the lesion (arrow) is well depicted as a 1.8-cm mass with moderately high signal intensity despite its small size. (b) In a patient with a hypovascular metastasis (arrow) from colon cancer, the low fluid content of the metastasis results in poor visibility of the lesion, which is a 4-cm mass with heterogeneous and minimally high signal intensity.
|
|
View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Coronal T2-weighted single-shot echo-train spin-echo MR images (/90) demonstrate fluid content of focal hepatic lesions. (a) In a patient with hemangioma, the high fluid content of the lesion (arrow) is well depicted as a 1.8-cm mass with moderately high signal intensity despite its small size. (b) In a patient with a hypovascular metastasis (arrow) from colon cancer, the low fluid content of the metastasis results in poor visibility of the lesion, which is a 4-cm mass with heterogeneous and minimally high signal intensity.
|
|
Various enhancement patterns in liver parenchyma are also demonstrated on hepatic arterial–dominant phase images. One of the most common perfusion abnormalities observed is transient increased segmental enhancement in liver segments with compromised portal venous flow due to compression or thrombosis (Fig 6) (24,25). Other hepatic diseases that demonstrate perfusion abnormalities on images obtained immediately after gadolinium enhancement include acute hepatocellular necrosis superimposed on chronic active hepatitis or cirrhosis; hepatic congestion; and Budd-Chiari Syndrome (26,27), with different enhancement patterns for acute, subacute, and chronic forms of the latter disease (27).
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. Transverse gadolinium-enhanced spoiled GRE MR images (150/4, 80° flip angle) obtained during the hepatic arterial-dominant phase in a patient with infected right main and left lateral segment portal venous thromboses and hepatic abscesses secondary to an appendiceal abscess. (a) Image obtained at the level of the right portal vein shows intense, early, transient, increased enhancement of the entire right lobe and regions of the left lateral segment, with heterogeneous and more normal enhancement of the medial segment of the left lobe, which reflects increased hepatic arterial delivery of contrast material to the segments with compromised portal venous supply. Thrombus (arrow) can be seen in the right portal vein. (b) At a more superior tomographic level (dome of the liver), a large abscess (arrow) with multiple internal septations is seen in the right lobe.
|
|
View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. Transverse gadolinium-enhanced spoiled GRE MR images (150/4, 80° flip angle) obtained during the hepatic arterial-dominant phase in a patient with infected right main and left lateral segment portal venous thromboses and hepatic abscesses secondary to an appendiceal abscess. (a) Image obtained at the level of the right portal vein shows intense, early, transient, increased enhancement of the entire right lobe and regions of the left lateral segment, with heterogeneous and more normal enhancement of the medial segment of the left lobe, which reflects increased hepatic arterial delivery of contrast material to the segments with compromised portal venous supply. Thrombus (arrow) can be seen in the right portal vein. (b) At a more superior tomographic level (dome of the liver), a large abscess (arrow) with multiple internal septations is seen in the right lobe.
|
|
Examination for liver metastases may be the most common indication for liver MR imaging. Liver metastases have been classified as hypovascular (typical examples are metastases from colon cancer and transitional cell carcinoma) or hypervascular (typical examples are metastases from islet cell tumors, renal cell carcinoma, and breast cancer) (2). A third category of vascularity that has not been described well is that of nearly isovascular with liver. Nearly isovascular refers to lesion enhancement that is comparable to that of liver on early- and late-phase gadolinium-enhanced images. Near isovascularity is most readily appreciated when lesions are poorly depicted on gadolinium-enhanced images but well depicted on nonenhanced images (Fig 7). Liver metastases from primary tumors of the colon, thyroid glands, and endometrium may demonstrate this type of enhancement pattern. The most common setting is in chemotherapy-treated disease, although this may also be observed without prior chemotherapy. Fortunately, many of these tumors have moderately low signal intensity on T1-weighted images, rendering them readily apparent; on occasion, they also may have moderately high signal intensity on T2-weighted images. Although gadolinium chelate administration may appear to be detrimental to patient diagnosis because lesions are poorly depicted on enhanced images, enhancement characteristics of nearly isovascular lesions in the presence of good lesion conspicuity on nonenhanced T1-weighted images is a feature of metastases and not of other lesions, and in the appropriate setting reflects a response to chemotherapy. On rare occasions, liver metastases may also be nearly isointense to liver parenchyma on both T1- and T2-weighted images and therefore can escape detection. The rarity of this event actually illustrates one of the great strengths of MR imaging over ultrasonography (US) and computed tomography (CT): namely, that the more different types of data acquired, the less likely it is for disease to escape detection. MR imaging simply can acquire more and different types of data than do US or CT.
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7a. Gadolinium-enhancement properties of nearly isovascular liver metastases. Transverse spoiled GRE MR images (160/4, 80° flip angle) obtained (a) before and (b) immediately after gadolinium enhancement show a 2.5-cm hypointense lesion (arrow in a) in the lateral segment of the left lobe of the liver. In b, the lesion has signal intensity that approaches that of the liver and is poorly demonstrated. This appearance of enhancement of a nearly isovascular lesion is a feature observed in chemotherapy-treated liver metastases.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7b. Gadolinium-enhancement properties of nearly isovascular liver metastases. Transverse spoiled GRE MR images (160/4, 80° flip angle) obtained (a) before and (b) immediately after gadolinium enhancement show a 2.5-cm hypointense lesion (arrow in a) in the lateral segment of the left lobe of the liver. In b, the lesion has signal intensity that approaches that of the liver and is poorly demonstrated. This appearance of enhancement of a nearly isovascular lesion is a feature observed in chemotherapy-treated liver metastases.
|
|
Chemotherapy-treated liver metastases deserve special mention because chemotherapy is routinely administered to many patients with liver metastases, and chemotherapy alters the imaging features of metastases. Chemotherapy induces changes in the signal intensity and imaging features of metastases and often results in alteration of signal intensity on both T1- and T2-weighted images, such that the signal intensity of metastases approaches that of liver (28,29). Chemotherapy-treated metastases may also develop an appearance that may resemble cysts, hemangiomas, or scar tissue (30). As mentioned earlier, they may also become nearly isovascular on gadolinium-enhanced images (Fig 7).
Portal venous phase or early hepatic venous phase.—Images for this phase are acquired at 45–60 seconds after initiation of the gadolinium chelate injection. In this phase, the hepatic parenchyma is maximally enhanced so that hypovascular lesions such as cysts, hypovascular metastases, and scar tissue are most clearly shown as regions of absent or diminished enhancement (Fig 8). Patency or thrombosis of hepatic vessels is also best shown during this phase (2). Peak hepatic parenchymal enhancement is achieved, but the characteristic enhancement features of many focal liver lesions frequently diminishes at this phase. Peak hepatic parenchymal enhancement is of greatest importance in the setting of hypovascular lesions, but this phase is often detrimental to the detection of hypervascular lesions such as hypervascular metastases and small hepatocellular carcinomas. Since colon cancer metastases are commonly observed in North America and the majority are hypovascular, much attention has been directed to their detection, which may benefit from image acquisition when the liver is maximally enhanced. MR imaging is more sensitive to the presence of gadolinium chelates than CT is to iodine-based contrast agents, and observation of ring enhancement of the outer most vascularized tumor margin is often appreciated even in hypovascular metastases. This may account for the greater importance of the hepatic arterial–dominant phase over the portal venous phase for liver lesion detection with MR imaging than has been generally recognized with CT.
View larger version (125K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. Transverse gadolinium-enhanced spoiled GRE MR images (160/4, 80° flip angle) of hypovascular liver metastases from a primary sarcoma. (a) During the hepatic arterial-dominant phase, the hypovascular metastasis is moderately visible (arrow). (b) During the portal venous phase, greater hepatic parenchymal enhancement improves the conspicuity of the metastasis (arrow), and ring enhancement is also more clearly defined.
|
|
View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. Transverse gadolinium-enhanced spoiled GRE MR images (160/4, 80° flip angle) of hypovascular liver metastases from a primary sarcoma. (a) During the hepatic arterial-dominant phase, the hypovascular metastasis is moderately visible (arrow). (b) During the portal venous phase, greater hepatic parenchymal enhancement improves the conspicuity of the metastasis (arrow), and ring enhancement is also more clearly defined.
|
|
Hepatic venous phase or interstitial phase.—This phase is acquired 90 seconds to 5 minutes after initiation of the contrast material injection. Late enhancement features of focal liver lesions are shown, which aid in lesion characterization. These features include persistent lack of enhancement of cysts, coalescence and centripetal progression of enhancing nodules in hemangiomas (15, 23), homogeneous "fading" of enhancement of adenomas and of focal nodular hyperplasia to near isointensity with liver parenchyma (31) (these lesions should never fade to hypointensity), late enhancement of central scar in some focal nodular hyperplasias (31), peripheral or heterogeneous washout of contrast material in liver metastases (32), washout to hypointensity relative to liver parenchyma in small hepatic metastases and hepatocellular carcinoma, heterogeneous washout of hepatocellular carcinoma, and delayed capsular enhancement in hepatocellular carcinoma (Fig 9) (less commonly in adenoma). Enhancement of peritoneal metastases; inflammatory disease; and circumferential, superficial, spreading cholangiocarcinoma are also well depicted during this time frame (10,33–35). Concomitant use of fat suppression is essential for optimal demonstration of these findings (Fig 10). Fat suppression diminishes the competing high signal intensity of fat to improve the conspicuity of enhancing diseased tissue. Fat suppression may also be used for the earlier gadolinium-enhanced MR sequences. In general, we have opted not to use fat suppression for these earlier sequences for the following reasons: (a) On hepatic arterial–dominant phase images, the most critical determination is detection and characterization of intraparenchymal lesions in the liver (and pancreas); (b) complete coverage of the liver is essential in the hepatic arterial dominant–phase, and fat suppression results in a reduction of the number of sections obtainable per data acquisition; and (c) the possible occurrence of heterogeneous fat suppression and, therefore, the reduction in the diagnostic quality of the hepatic arterial–dominant phase images are disadvantageous. In addition, documentation of vascular thrombosis is provided on these images.
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. Small hepatocellular carcinoma demonstrated on transverse gadolinium-enhanced (a) hepatic arterial-dominant phase spoiled GRE (160/4, 80° flip angle) and (b) interstitial phase fat-suppressed spoiled GRE (170/4, 80° flip angle) MR images. (a) Transient intense tumor blush (arrow) is appreciated on the hepatic arterial-dominant phase image. (b) Tumor signal intensity has faded to lower than that of background liver on the interstitial phase image, and delayed enhancement of the tumor capsule can be appreciated. Early tumor blush, tumor washout, and delayed capsular enhancement are features of hepatocellular carcinoma.
|
|
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. Small hepatocellular carcinoma demonstrated on transverse gadolinium-enhanced (a) hepatic arterial-dominant phase spoiled GRE (160/4, 80° flip angle) and (b) interstitial phase fat-suppressed spoiled GRE (170/4, 80° flip angle) MR images. (a) Transient intense tumor blush (arrow) is appreciated on the hepatic arterial-dominant phase image. (b) Tumor signal intensity has faded to lower than that of background liver on the interstitial phase image, and delayed enhancement of the tumor capsule can be appreciated. Early tumor blush, tumor washout, and delayed capsular enhancement are features of hepatocellular carcinoma.
|
|
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10. Transverse interstitial phase gadolinium-enhanced fat-suppressed spoiled GRE MR image (170/4, 80° flip angle) shows peritoneal metastases (arrows) from endometrial stromal sarcoma. The high-signal-intensity metastases are well depicted against a background of low-signal-intensity ascites and suppressed fat.
|
|
Comparison with CT
The preferential use of MR imaging over CT in the investigation of liver disease cannot be justified unless the advantages of MR imaging are well documented. Clearly, the lack of ionizing radiation with MR imaging and the safety of gadolinium chelates, as compared with that of iodinated contrast agents, are two important considerations. With regard to the safety of gadolinium, it has been shown that gadolinium chelates have negligible effects on renal function in patients with renal failure (36), unlike iodinated agents, which have a deleterious effect. Gadolinium chelates are also much less likely to result in major allergic reactions than even nonionic iodinated contrast agents. At many institutions, therefore, patients with an elevated serum creatinine level or a history of allergy to iodinated contrast agents routinely undergo MR imaging to investigate for focal liver disease, and these conditions are a major indication for MR imaging.
In addition to safety issues, the authors of virtually all reports since 1990 (6–9,12–14,37–41) in which current MR techniques have been compared with current CT approaches have shown that MR imaging is more accurate than iodinated contrast material–enhanced CT for the detection and characterization of liver lesions. Although both modalities have evolved throughout this period, MR imaging has remained superior for this application. Earlier studies compared MR imaging with dynamic contrast-enhanced CT (6,7,9), whereas later studies compared MR imaging with CT during arterial portography (38,40) or spiral CT (8,12,13,41). The more recent comparisons have been between hepatic arterial–dominant phase gadolinium-enhanced MR imaging and hepatic phase iodinated contrast-enhanced CT (12,13). The relative paucity of these comparative studies is unfortunate, despite the importance of determining the optimal diagnostic investigation method.
Extrahepatic Abdominal Investigation
Although investigation of liver disease may be the most common indication for abdominal imaging, for an imaging modality to be widely utilized it must be accurate in the investigation of organs and tissues in addition to the liver. Few studies have evaluated the MR imaging demonstration of abdominal disease outside the liver, and few have compared MR imaging with CT, the accepted imaging standard for extrahepatic disease, to demonstrate the role for MR imaging (10,42,43). Recently Low et al (10) described a prospective blinded comparison between MR imaging and spiral CT for the evaluation of extrahepatic disease. They found that MR imaging during the hepatic arterial–dominant phase with gadolinium-enhanced and hepatic venous phase fat-suppressed spoiled GRE sequences demonstrated significantly more surgically-proved sites of extrahepatic disease than did spiral CT. They found that CT demonstrated 101 (65%) and MR imaging 140 (90%) of 155 surgically proved sites (P < .001).
RES-specific Contrast Agents
Iron oxide particulate agents are selectively taken up by RES cells in the liver, spleen, and bone marrow and result in signal loss at T2-weighted imaging owing to the susceptibility effects of iron (44). This class of contrast agent is also termed superparamagnetic iron oxide, and the first of these agents licensed for use in the United States is ferumoxides (Feridex; Advanced Magnetics, Cambridge, Mass) (44). Lesions that contain negligible or few RES cells remain largely nonenhanced, while the normal liver enhances (ie, becomes low signal intensity on T2-weighted images), with the result that the contrast-to-noise ratio between enhanced (low-signal-intensity) liver and nonenhanced (persistently high-signal-intensity) liver lesions is improved on ferumoxides-enhanced T2-weighted images, as compared with nonenhanced T2-weighted images (40,41,44–46). This affords both increased lesion conspicuity and increased lesion detection, as compared with nonenhanced images (Fig 11).
View larger version (120K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Ferumoxides-enhanced MR images of liver metastases. (a) Transverse nonenhanced T2-weighted fat-suppressed echo-train spin-echo MR image (3,500/90) shows two mildly hyperintense focal liver lesions (arrows). (b) Transverse ferumoxides-enhanced T2-weighted fat-suppressed echo-train spin-echo MR image (3,500/90) show that the signal intensity of liver, spleen, and bone marrow are diminished, which renders the liver metastases more conspicuous. Additional small (<1-cm-diameter) metastases (arrows) are demonstrated that were not apparent in a.
|
|
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Ferumoxides-enhanced MR images of liver metastases. (a) Transverse nonenhanced T2-weighted fat-suppressed echo-train spin-echo MR image (3,500/90) shows two mildly hyperintense focal liver lesions (arrows). (b) Transverse ferumoxides-enhanced T2-weighted fat-suppressed echo-train spin-echo MR image (3,500/90) show that the signal intensity of liver, spleen, and bone marrow are diminished, which renders the liver metastases more conspicuous. Additional small (<1-cm-diameter) metastases (arrows) are demonstrated that were not apparent in a.
|
|
The type of lesion for which improved detection with this contrast agent may be most clinically useful is liver metastases (46). Studies have shown that ferumoxides-enhanced T2-weighted MR imaging has performance comparable to that of CT during arterial portography for the demonstration of liver metastases (40), which suggests that the role for this contrast agent may be as a replacement for CT during arterial portography as a modality to improve detection of liver metastases in patients already known to have liver metastases. The patient group in which this role for ferumoxides may be the most applicable is patients with liver metastases from colon cancer who are considered to be candidates for hepatic resection on the basis of the suspicion of limited involvement of the liver with metastatic disease (46).
Although serious adverse events are rare, approximately 3% of patients will experience severe back pain while the contrast agent is being administered (44). This back pain appears to be a side effect of particulate agents in general; it is not specific to ferumoxides and is limited to the injection period and slightly beyond. Back pain develops in patients in whom the contrast agent is administered too rapidly (ie, faster than the recommended slow intravenous drip infusion over 30 minutes) and is more likely to occur in patients with liver dysfunction, such as patients with cirrhosis. Reduction of the injection rate or termination of the injection, allowing the back pain to resolve, and reinitiation of the administration at a slower rate are generally sufficient measures, without loss of diagnostic quality of the images. It is important to realize that it is not necessary to terminate the examination.
The long infusion period (30 minutes) is an inconvenient aspect of this agent, which necessitates two imaging sessions for nonenhanced and enhanced images. Attractive features of the agent include the long imaging window (1–4 hours), no need for precise dynamic image acquisition related to contrast material administration (unlike with nonspecific extracellular gadolinium chelates), and acceptable image quality with machines of various field strength with no requirement for patient breath holding or that the MR machine be capable of performing breath-hold sequences (44). These latter two requirements are necessary with nonspecific extracellular gadolinium chelates (6–8).
Many centers use breathing-averaged T2-weighted fat-suppressed sequences in conjunction with this contrast agent. Image quality in general is more reproducible with breath-hold or breathing-independent T2-weighted sequences. Breath-hold echo-train spin-echo, breath-hold short-inversion-time inversion recovery, breath-hold GRE, and breathing-independent echo-train spin-echo sequences are all effective with this agent. The echo-train sequences do, however, result in a decrease in the T2* effects of the agent due to the narrow echo spacing. Unlike the case with T1-weighted breath-hold spoiled GRE sequences, used for gadolinium-enhanced studies, the imaging parameters that are optimal for use with this agent have not been established. Empirically, one sequence that is effective is a GRE sequence at 1.5 T with a TR of 150 msec, a TE of 9 msec, and a flip angle of 45°; these parameters were so chosen to utilize the T2* effects of the contrast agent.
A cautionary note is that the dose of Feridex that is licensed for use in North America is approximately one-third less than that used in Europe. The importance of this is that sequences effective with one dose may not be effective with another. For example, a T1-weighted spoiled GRE sequence, which is essentially the same sequence that we use with nonspecific extracellular gadolinium chelates, has been reported as an ideal sequence for use with ferumoxides in a European study (45), whereas the same sequence with the dose of contrast material used in North America will generally result in obscuring of the lesions. The explanation for this is that the dose of Feridex in North America is not enough to produce sufficient T2* effects to lower the signal intensity of liver below that of liver lesions such as metastases in a consistent reproducible fashion.
The increased uptake of ferumoxides agents by focal nodular hyperplasia, because of its high content of RES cells, has also been investigated as a possible clinical role for ferumoxides. There may, however, be too much of an overlap of the concentration of RES cells in focal nodular hyperplasia, adenoma, and hepatocellular carcinoma for this clinical application to be sufficiently accurate for routine use.
Iron oxide particles have also been formulated as smaller particulate agents, which have been termed ultrasmall particulate iron oxides (47). These agents have a more prolonged intravascular half-life than do the larger particle agents, and in a dilute intravascular phase they possess T1 effects that emulate the vascular phase effects of T1 agents. Therefore, they can provide additional characterization information, such as the demonstration of peripheral nodular enhancement in hemangiomas. These agents also can provide bright vessel enhancement in the vascular phase, which can be used for MR angiography. Another major advantage of the ultrasmall iron particle agents over the small iron particle agents is that they can be administered as a rapid intravenous bolus in a small volume. They are more convenient to use and, because of the small dose, should not result in back pain as an adverse effect.
Hepatocyte-selective Contrast Agents
Hepatocyte-selective contrast agents undergo uptake by hepatocytes and are eliminated through the biliary system. This category of contrast agents comprises T1 agents, which result in increased signal intensity on T1-weighted images of tissues that show contrast material uptake. Normal liver and focal hepatic lesions that contain hepatocytes take up these agents, and lesions that do not contain hepatocytes do not take up these agents (48,49).
Currently, the only hepatocyte-selective contrast agent that is licensed for use in the United States is mangafodipir trisodium, formerly known as Mn-DPDP (48). This agent is administered as a slow (1-minute) intravenous infusion, and the maximal imaging window is between 15 minutes and 4 hours. Dynamic images are not acquired with this agent, but virtually all T1-weighted sequences can be used. Unlike with nonspecific extracellular gadolinium chelates, adequate imaging with mangafodipir trisodium does not require that the MR machine have high field strength or that breath-hold sequences be performed, nor does it require that the patient be able to hold his or her breath. Because dynamic imaging is not essential, sequences that produce a higher signal-to-noise ratio or higher spatial resolution and that may not acquire sufficient sections to encompass the entire liver in one data acquisition can be used effectively. This is not the case with the nonspecific extracellular gadolinium chelates, with which, as previously stated, the best results are achieved with complete liver coverage during one breath hold. A useful sequence is spoiled GRE with a matrix size of 512 x 512 to achieve higher spatial resolution with this contrast agent (Fig 12). At this early stage of clinical use, this agent appears to be safe and well tolerated, despite the fact that free manganese may be the active contrast agent.
View larger version (121K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12a. T1-weighted mangafodipir trisodium-enhanced spoiled GRE MR images (180/4, 80° flip angle, 512 x 512 matrix) of liver metastases. (a) Coronal and (b) transverse high-spatial-resolution images demonstrate fine detail of liver metastases (large arrow). Biliary excretion of manganese is shown as high-signal-intensity fluid (small arrow) in the biliary system.
|
|
View larger version (147K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12b. T1-weighted mangafodipir trisodium-enhanced spoiled GRE MR images (180/4, 80° flip angle, 512 x 512 matrix) of liver metastases. (a) Coronal and (b) transverse high-spatial-resolution images demonstrate fine detail of liver metastases (large arrow). Biliary excretion of manganese is shown as high-signal-intensity fluid (small arrow) in the biliary system.
|
|
At present, the best clinical role for mangafodipir trisodium appears to be as an agent to improve detection of the number and extent of focal liver lesions, particularly liver metastases from colon cancer, in patients in whom hepatic resection is being contemplated (Fig 12). As with ferumoxides, the tissue-selective aspects of mangafodipir trisodium are under investigation (48,49). Focal liver lesions such focal nodular hyperplasia, adenoma, and well-differentiated hepatocellular carcinoma all take up the agent, whereas hemangiomas and metastases do not (48). It is not clear how often the distinction between hepatocyte-containing and non–hepatocyte-containing lesions is of clinical importance, since whether a lesion is benign or malignant is often the more important clinical question.
A feature of hepatocyte-selective contrast agents is that they are eliminated in part through the biliary system. They therefore permit evaluation of hepatocyte function and of the appearance and integrity of the biliary tree (50).
Contrast Agents with Combined Perfusion and Hepatocyte-selective Properties
These contrast agents exhibit elimination through both renal and hepatic excretion pathways and thereby possess both early perfusion information (renal elimination pathway) and later hepatocyte-selective information (hepatic excretion pathway). These agents therefore combine the liver lesion detection and characterization information provided by nonspecific extracellular gadolinium chelates with the liver lesion detection and characterization (hepatocyte vs nonhepatocyte) information provided by hepatocyte-selective contrast agents (Fig 13). Gadolinium benzyloxypropionictetraacetate, or Gd-BOPTA (51), and gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid, or Gd-EOB-DTPA (52,53), are examples of this category of contrast agent. This category of contrast agent may prove to be the most effective for evaluating liver disease.
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13a. Transverse spoiled GRE images (170/4, 80° flip angle) of metastases obtained (a) immediately after and (b) 10 minutes after administration of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid. (a) Immediately after contrast material administration, perfusion information is provided, with demonstration of ring enhancement (arrow) around the metastasis. (b) Ten minutes later, hepatocellular uptake is shown as increased signal intensity of the hepatic parenchyma, while the metastasis (arrow) appears as a well-defined hypointense lesion.
|
|
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 13b. Transverse spoiled GRE images (170/4, 80° flip angle) of metastases obtained (a) immediately after and (b) 10 minutes after administration of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid. (a) Immediately after contrast material administration, perfusion information is provided, with demonstration of ring enhancement (arrow) around the metastasis. (b) Ten minutes later, hepatocellular uptake is shown as increased signal intensity of the hepatic parenchyma, while the metastasis (arrow) appears as a well-defined hypointense lesion.
|
|
Comparison between Contrast Agents
Several authors (6–9,12,13,37–41) have shown that MR imaging enhanced with various contrast agents is superior to various forms of state-of-the-art CT. On the basis of these study results, it would appear reasonable to believe that contrast-enhanced MR imaging is superior to CT for the detection of focal liver lesions. There is, however, a paucity of studies that have attempted to compare MR contrast agents among one another (46,53–55). Because nonspecific extracellular gadolinium chelates are not only relatively inexpensive but also provide comprehensive information on a wide range of abdominal diseases (10), definite advantages of newer agents or specific roles for new agents must be devised. Table 2 provides a comparison between various agents with regard to their contrast-enhancement mechanism, target tissue, and clinical properties.
View this table:
[in this window]
[in a new window]
|
TABLE 2. Comparison of Contrast-Enhancement Mechanisms, Target Tissues, and Clinical Properties of MR Contrast Agents
|
|
|
SUMMARY
|
|---|
Various categories of MR contrast agents are available for clinical use, all of which have been shown to enable the demonstration of more liver lesions than can be depicted on nonenhanced MR images and images obtained with current CT techniques. Nonspecific extracellular gadolinium chelates are safe, relatively inexpensive, and enable acquisition of information on detection and characterization of a wide range of focal liver lesions and other liver diseases, as well as information on the extrahepatic abdomen. Dynamic image acquisition is important with these agents, and correct timing of data acquisition to capture the hepatic arterial–dominant phase is critical.
The clinical application of new contrast agents must evolve into an integrated diagnostic scheme to supplement information provided with the use of nonspecific extracellular gadolinium chelates (eg, detection of additional metastases in patients with known metastases who are considered for surgical resection) or in patient groups in whom images obtained with nonspecific extracellular gadolinium chelate enhancement may not show lesions well (eg, postchemotherapy images) or when they provide novel information (eg, demonstration of biliary elimination of hepatocyte selective agents).
|
FOOTNOTES
|
|---|
Abbreviations: GRE = gradient echo,
RES = reticuloendothelial system,
TE = echo time,
TR = repetition time
R.C.S. is on the speakers bureaus for Berlex (Wayne, NJ) and Nycomed Amersham (Princeton, NJ).
|
REFERENCES
|
|---|
TOP
ABSTRACT
INTRODUCTION
