| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of Vienna, Waehringer Gürtel 18-20,
A-1090 Vienna, Austria.
2 Department of Radiology, University of Washington, Harborview Medical Center,
Box 359728, Seattle, WA 98104-2499.
3 Department of Radiology, Klinikum Karlsruhe, Moltkestr. 90, D-76133 Karlsruhe,
Germany.
4 Department of Surgery, University of Vienna, A-1090 Vienna, Austria.
5 Department of Pathology, University of Vienna, A-1090 Vienna, Austria.
6 Department of Nuclear Medicine, University of Vienna, A-1090 Vienna,
Austria.
7 Department of Radiology, Duke University Medical Center, Durham, NC
27710.
Received January 7, 2002;
accepted after revision June 10, 2002.
Presented at the annual meeting of the American Roentgen Ray Society,
Seattle, April—May 2001.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Contrast-enhanced MR images of 45 patients with 85 focal nodular hyperplasia lesions were retrospectively reviewed. In these patients, sonographic findings were nonspecific (n = 37), or CT features were inconclusive (n = 8). Non—liver specific gadolinium chelates were used in 18 patients (48 lesions) suspected of having either focal nodular hyperplasia or hemangioma. The following liver-specific agents were used in patients with suspected focal nodular hyperplasia or metastases: mangafodipir trisodium, 30 patients (55 lesions); ferumoxides, six patients (16 lesions); and SHU 555 A, six patients (six lesions). Individual lesions were quantified by signal intensity and assessed qualitatively by homogeneity, contrast enhancement, and presence of a central scar.
RESULTS. At unenhanced MR imaging, the triad of homogeneity, isointensity, and central scar was found in 22% of the focal nodular hyperplasia lesions. On mangafodipir trisodium—enhanced T1-weighted images, all focal nodular hyperplasia lesions showed contrast uptake: in 64% of the lesions, uptake was equal to parenchyma; 25%, greater than the parenchyma; and 11%, less than the parenchyma. On iron oxide—enhanced T2-weighted images, all focal nodular hyperplasia lesions showed uptake of the contrast agent, but contrast uptake in the lesions was less than in the surrounding parenchyma. Dynamic gadolinium chelate—enhanced MR imaging showed early and vigorous enhancement of focal nodular hyperplasia lesions with rapid washout in 88%. Atypical imaging features of the lesions included hyperintensity on T1-weighted images, necrosis and hemorrhage, and inhomogeneous or only minimal contrast uptake.
CONCLUSION. For patients in whom the diagnosis of focal nodular hyperplasia cannot be established on unenhanced or gadolinium-enhanced MR imaging, homogeneous uptake of liver-specific contrast agent with better delineation of central scar may help to make a confident diagnosis of focal nodular hyperplasia.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Dynamic gadolinium-enhanced MR imaging enables evaluation of the fast hemokinetics of the vascular—interstitial contrast distribution during the first few minutes after injection [7,8,9]. Although focal nodular hyperplasia lesions have been reported to show rapid enhancement during the arterial phase and slow, gradual enhancement of the central scar during the delayed phase [7,8,9,10,11], a similar enhancement pattern during the arterial phase is also seen in hypervascular lesions other than focal nodular hyperplasia, such as hypervascular metastasis (e.g., renal cell carcinoma, melanoma, islet cell tumor), adenoma, and classic or fibrolamellar hepatocellular carcinoma [9, 10].
MR imaging contrast agents that are selectively accumulated in specific liver cell types (reticuloendothelial, hepatobiliary) provide direct insight into cellular constituency of individual lesions [12, 13]. In particular, Kupffer's cells in focal nodular hyperplasia lesions may sequester a detectable amount of superparamagnetic iron oxide particles, such as ferumoxides and SHU 555 A [13,14,15,16,17]. However, well-differentiated hepatocellular carcinomas and hepatic adenomas may show avid uptake of superparamagnetic iron oxide [16]. The use of mangafodipir trisodium, a contrast agent taken up by hepatocytes and eliminated through the biliary system of non-hepatocellular tumors, improves detection [12, 18]. Although focal nodular hyperplasia lesions will show uptake, so too may many hepatic adenomas and hepatocellular carcinomas [18, 19].
For many liver lesions, no single MR imaging strategy can sufficiently characterize abnormalities to avoid biopsy. We thus hypothesize that the combination of morphologic, hemokinetics, and cellular composition information derived from unenhanced imaging, dynamic vascular enhanced imaging, and the administration of cell-specific contrast agents, respectively, may be helpful in characterizing focal nodular hyperplasia on MR images.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mean size of all 85 lesions was 4.5 cm (range, 1.5-12.0 cm). Sixty-one lesions were located in the right lobe, 13 in the left lobe, and 11 in the caudate lobe. A single lesion was depicted in 27 patients, and multiple lesions were present in the other 18 patients (two lesions in seven patients, three in five patients, four in one patient, and five in five patients).
Histologic proof of focal nodular hyperplasia was obtained in 58 lesions (35 patients). Percutaneous needle biopsy was performed in 24 patients with 37 lesions. Surgical resection was performed in 11 patients with 21 subcapsular or large exophytic lesions because of symptoms such as discomfort, abdominal pain, or compression of the inferior vein cava. In the 27 lesions not confirmed at histology (10 patients), typical radiologic features were shown at follow-up and the appearance remained unchanged on MR imaging over at least 24 months (range, 24-33 months).
MR Imaging
The MR examinations were performed using a 1.5-T unit (ACS-NT, Philips
Medical Systems, Best, The Netherlands [15 patients]; or Vision, Siemens,
Erlangen, Germany [30 patients]). Unenhanced T1-weighted gradient-recalled
echo in- and opposed-phase sequences and T2-weighted turbo spin-echo sequences
with fat saturation were performed in all patients. In addition,
T2*-weighted gradient-recalled echo images were obtained in 12
patients.
The examination parameters were as follows when using the Vision MR unit. Fast low-angle shot imaging was performed for the T1-weighted gradient-recalled echo in-phase and opposed-phase sequences. For the in-phase sequence, the TR/TE was 177/4 and the flip angle was 80°; for the opposed-phase sequence, 133/2 and 70°. When performing the T2-weighted fat-saturated turbo spin-echo sequence, we used 4000/88 and an echo-train length of 33. With the ACS-NT unit, a T1-weighted gradient-recalled echo sequence was performed with 14/4 and a flip angle of 30°, and a fat-saturated T2-weighted turbo spin-echo sequence was performed with 3000/100 and an echo-train length of 23. The slice thickness was 8-10 mm, and the interslice gap was 0.8-1.
The following contrast agents were used. Non—liver specific gadolinium chelates (Magnevist [gadopentetate dimeglumine], Schering, Berlin, Germany; Omniscan [gadodiamide hydrate], Amersham Health, Oslo, Norway) were administered in 18 patients (48 lesions). Mangafodipir trisodium (Teslascan; Amersham Health) was given to 30 patients (55 lesions). Ferumoxides (endorem; Guerbet, Roissy, France) were administered in six patients (16 lesions), and SHU 555 A (Resovist; Schering) was used in six patients (six lesions).
Only one contrast agent was given during a single MR examination. The time interval between the injection of the first MR contrast agent and the second one, if performed, ranged from 3 days to 6 months. After manual IV bolus administration of non—liver specific gadolinium chelates, T1-weighted dynamic gradient-recalled echo sequences were performed during the arterial dominant phase (20 sec after injection), the portal venous phase (60 sec after injection), and the delayed phase (4 min after injection).
After IV infusion of mangafodipir trisodium, T1-weighted gradient-recalled echo images were obtained 20 min after start of the infusion. After IV infusion of ferumoxides over 30 min or IV bolus application of SHU 555 A, T2-weighted fatsaturated turbo spin-echo and T2*-weighted gradient-recalled echo sequences were performed.
The choice of MR contrast agents depended on sonographic and CT findings: When hemangiomas were suspected, contrast-enhanced MR imaging was performed using non—liver specific gadolinium chelates. When metastasis, cholangiocarcinoma, or focal nodular hyperplasia was suspected, patients received liver-specific MR contrast agents (i.e., mangafodipir trisodium or iron oxide particles).
Image Analysis
The MR images were qualitatively and quantitatively assessed. The
qualitative analysis of liver lesions was performed by two radiologists in
consensus and with the knowledge of the diagnosis of focal nodular
hyperplasia. The following MR features were assessed: signal intensity of
lesions compared with normal parenchyma on T1- and T2-weighted images; lesion
homogeneity; presence or absence of a central scar; and, after administration
of contrast agents, lesion enhancement on T1-weighted images or signal
intensity loss, predominately on T2-weighted images.
For quantitative image analysis, the signal intensity of focal nodular
hyperplasia and of liver parenchyma and the background noise were measured
using operator-defined regions of interest. The largest possible region of
interest in the lesion that excluded necrotic and fibrotic areas was selected
for measurement of signal intensity by one radiologist. A region of interest
greater than 4 cm2 that excluded vessels and artifacts was used to
measure the signal intensity of the liver adjacent to the tumor. The regions
of interest were placed identically for both the unenhanced and
contrast-enhanced images. Background noise was measured anterior to the liver
in the phase-encoding direction. To avoid volume averaging, the reviewer
quantitatively assessed only lesions larger than 1 cm. The lesion-to-liver
contrast-to-noise ratio was calculated as the difference in signal intensity
between the lesion and the liver scaled to the standard deviation of
background noise. The percentage of contrast enhancement was calculated for
gadolinium chelates and mangafodipir trisodium as follows:
 |
 |
The Student's t test for paired samples was used to compare the contrast-to-noise ratios for the various pulse sequences before and after injection of nonspecific or liver-specific contrast agents.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The results concerning the qualitative assessment of focal nodular hyperplasia on contrast-enhanced MR images are shown in Table 1. In 48 of the 85 lesions, gadolinium chelates were IV injected. Forty-two of these lesions (88%) were hyperintense to the liver during the arterial dominant phase (20-30 sec after bolus injection) (Figs. 1A,1B,1C,1D,1E and 3A,3B,3C,3D,3E), five focal nodular hyperplasia lesions (10%) became isointense relative to the liver (Figs. 2A,2B,2C,2D,2E and 4A,4B,4C,4D,4E), and one lesion remained hypointense (2%) on T1-weighted gradient-recalled echo images. Enhancement of the central scar, which became hyperintense, was seen in 26 lesions (54%) within 60-240 sec after the injection ended (Figs. 1A,1B,1C,1D,1E and 3A,3B,3C,3D,3E). Fourteen lesions that were isointense on unenhanced images and were not depicted could be easily seen on the arterial dominant gadolinium-enhanced images (Fig. 3A,3B,3C,3D,3E). Furthermore, three additional pseudocapsules and one central scar were depicted, and five central scars and one pseudocapsule were better visualized.
After mangafodipir trisodium injection, 55 focal nodular hyperplasia nodules were enhanced on T1-weighted images (Table 2). Thirty-five lesions were isointense to the surrounding liver parenchyma (Figs. 1A,1B,1C,1D,1E and 3A,3B,3C,3D,3E), 14 were hyperintense (Fig. 4A,4B,4C,4D,4E), and six were slightly hypointense to the surrounding liver parenchyma. On mangafodipir trisodium—enhanced T1-weighted images, no enhancement of the central scars was depicted. In four focal nodular hyperplasia lesions, the scars and pseudocapsules were shown only on the mangafodipir trisodium—enhanced T1-weighted images. Thus, the hypointense scar and septa were better visualized on contrast-enhanced T1-weighted images (Fig. 1A,1B,1C,1D,1E).
|
After administration of ferumoxides or SHU 555 A, all 22 focal nodular hyperplasia nodules were slightly hyperintense (Figs. 2A,2B,2C,2D,2E and 4A,4B,4C,4D,4E). The scars and septa were better visualized on T2-weighted images because the contrast between the scar and the tumor was increased. On unenhanced images, a central scar was visible in 11 lesions.
After administration of superparamagnetic iron oxide particles, a central scar was revealed in five additional lesions. In a patient who underwent follow-up studies with three different contrast agents, ferumoxides and mangafodipir trisodium revealed two additional lesions that had been isointense on gadolinium-enhanced images and could not be seen.
The results of the quantitative analysis are summarized in Table 3. Regarding the quantitative analysis, gadolinium-enhanced T1-weighted gradient-recalled echo images had the best contrast-to-noise ratio (35.7 ± 5.8) (p < 0.05). The signal-to-noise ratio of focal nodular hyperplasia lesions was significantly improved on the mangafodipir trisodium—enhanced gradient-recalled echo T1-weighted images compared with unenhanced images (p < 0.05). The contrast-to-noise ratio of focal nodular hyperplasia lesions was slightly improved on mangafodipir trisodium—enhanced T1-weighted images compared with the unenhanced images. However, the difference did not reach a statistical significance (p = 0.221) because most lesions became isointense. Superparamagnetic iron oxide—enhanced T2-weighted images showed a decrease in the signal-to-noise ratio of focal nodular hyperplasia lesions (p < 0.05). However, the decrease in signal intensity of the focal nodular hyperplasia lesions was less than the decrease in the signal intensity of the liver on all sequences, so the lesions were easier to detect. The signal intensity loss on contrast-enhanced images and the increase in lesion-to-liver contrast-to-noise ratio on T2-weighted images after infusion of superparamagnetic iron oxide was statistically significant (p < 0.05). Although there was a large range in the size of the lesions included in the study, no significant difference in the enhancement pattern was detected.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The wide range of MR appearances of focal nodular hyperplasia lesions on unenhanced MR images can be explained in part by patient selection. In series with a large number of patients undergoing surgery, a large number of lesions with atypical characteristics will be observed because the diagnosis could not be made using imaging findings alone. On the other hand, the use of different magnetic field strengths or different MR sequences may account for the variety of imaging features reported [8, 20].
A pseudocapsule has also been reported as an atypical feature of focal nodular hyperplasia [21]. We detected a pseudocapsule in 9% of the lesions on unenhanced images and in 18% of the lesions on contrast-enhanced images of our series. Dilated vessels and sinusoids around focal nodular hyperplasia lesions are well-documented sources of a pseudocapsule [2, 4]. Thus, a pseudocapsule should not be regarded as an atypical feature.
Previous studies [20, 22] reported that only 2.1% of focal nodular hyperplasia lesions were hyperintense on T1-weighted images. Hyper-intensity of these lesions on T1-weighted images can be caused by various pathologic changes including fat deposition, copper accumulation, high protein concentration, blood degradation products, or sinusoidal dilatation [22]. Sinusoidal dilatations were present histologically within the lesions in our study. Fat saturation is helpful in the detection and differentiation of fat deposits in hyperintense lesions on T1-weighted images [22]. In our study, all seven lesions that appeared hyperintense on T1-weighted images remained hyperintense on fat-suppressed images.
MR imaging provides useful strategies for detecting and characterizing liver tumors [4]. However, because of wide biologic variability and considerable overlap in the T1 and T2 relaxation times of tumor and normal liver, many lesions exhibit only a subtle change in intensity compared with normal liver tissue or are isointense to normal liver tissue on unenhanced images. Recognition of small lesions on unenhanced images can be difficult because of the relative isointensity of the lesion to the liver. After the administration of contrast agents, characterization of focal liver lesions can be improved by assessing the spatial distribution of contrast material uptake and the temporal pattern of contrast material uptake or washout [8,9,10,11].
In our series, gadolinium-enhanced images revealed 14 additional focal nodular hyperplasia lesions (29%) that were not found on unenhanced MR images. We observed one central scar and three pseudocapsules after administration of gadolinium chelates. Furthermore, better visualization of five central scars allowed a more confident diagnosis. Our findings indicate that characterization of focal nodular hyperplasia is facilitated by dynamic MR imaging after IV administration of gadolinium chelates.
MR imaging diagnosis of focal nodular hyperplasia with the non—liver specific gadolinium chelates is based on the early hyperintensity of the lesion and delayed contrast enhancement of the central scar because the vascular supply to a focal nodular hyperplasia lesion is predominantly arterial [8,9,10,11]. However, other hepatocellular tumors, such as adenomas or carcinomas, and other types of hypervascular tumors, such as cholangio-carcinoma and metastases (e.g., carcinoid, melanoma, islet cell tumor, or renal cell carcinoma), can display similar features. Thus, the diagnosis of focal nodular hyperplasia is sometimes difficult to establish [8, 10, 23,24,25]. In our study, 88% of the focal nodular hyperplasia nodules showed strong, homogeneous enhancement on gadolinium-enhanced images during the arterial dominant phase, and 12% of the lesions showed only modest or minimal contrast enhancement, which is not typical for focal nodular hyperplasia.
The interest in liver-specific contrast agents continues because findings in clinical studies show that contrast-enhanced MR imaging performed with non—liver specific extracellular contrast agents does not always lead to improved detection and characterization of liver tumors, and overlapping enhancement patterns may cause difficulties in tissue characterization [8,9,10]. When liver-specific MR contrast agents are used, functional and morphologic features due to the presence of Kupffer's cells or hepatocytes in focal nodular hyperplasia lesions reveal additional enhancement characteristics of focal nodular hyperplasia [12,13,14,15,16,17,18,19, 26, 27].
With the use of different MR contrast agents, one type of tissue should show high uptake of a certain contrast agent, whereas another type of tissue should remain entirely free of this contrast agent; however, this ideal situation is not always achieved. Liver-specific contrast agents, which show uptake only in liver lesions that are composed of functional hepatocytes or Kupffer's cells, allow differentiation between metastatic liver diseases and hepatocellular tumors [16, 19]. There are two classes of liver-specific agents. The hepatobiliary agents, which are predominantly used for T1-weighted imaging, are taken up into the hepatocytes and excreted into the biliary tract. The reticuloendothelial agents (or Kupffer's cell agents) are sequestered by the reticuloendothelial system and are deposited mainly in the Kupffer's cells of the liver and in the spleen [27]. Because focal nodular hyperplasia lesions contain a reticuloendothelial system and a lower density of hepatocytes relative to normal liver parenchyma, contrast between the lesion and surrounding tissue is increased due to the significant loss of signal intensity in normal liver tissue.
The appearance of focal nodular hyperplasia lesions on MR imaging after the administration of superparamagnetic iron oxide particles has been described [13,14,15,16,17]. Ferumoxides and SHU 555 A are phagocytosed by Kupffer's cells in focal nodular hyperplasia lesions, just as in normal liver parenchyma. The decrease in signal intensity of the lesion, however, may be less than that of normal liver because of the variable content of Kupffer's cells. Our results are concordant with previous studies [14,15,16, 26] that revealed marked uptake of contrast agent into focal nodular hyperplasia nodules and marked signal intensity loss in the lesions (approximately 30%) but less than that in the surrounding liver tissue (>68%). Therefore, the lesions and the central scar appear more conspicuous when these contrast agents are used. Furthermore in our study, two additional focal nodular hyperplasia lesions were shown, and five additional central scars were depicted after the administration of superparamagnetic iron oxide particles. Visualization of all these features led to a more accurate diagnosis.
The hepatobiliary liver-specific contrast agent mangafodipir trisodium is taken up by hepatocytes and eliminated through the biliary system [12, 18, 19]. This contrast agent has been reported to be taken up by focal nodular hyperplasia [19]. In our study, the signal intensity of focal nodular hyperplasia lesions was equal to (64% of the lesions) or slightly lower than (11% of the lesions) the signal intensity of normal liver on mangafodipir trisodium—enhanced images in most cases. However, in 25% of the cases, the signal intensity of the focal nodular hyperplasia lesion was higher than that of the surrounding normal liver on MR images obtained after contrast infusion. This difference in intensity may seem surprising because both liver and focal nodular hyperplasia consist of normal hepatocytes, but it can be explained by a higher rate of contrast uptake or a lower rate of elimination of the contrast agent in focal nodular hyperplasia lesions compared with normal liver (or both) [19].
Although the scar of focal nodular hyperplasia is known to be rich in bile ducts [28], our study showed no enhancement of the scar with mangafodipir trisodium. This lack of enhancement may be attributed to the bile ducts—their failure to function, their small size, their architectural disorganization, or their compression by the fibrous tissue [19]. The administration of mangafodipir trisodium showed two additional lesions that could not be seen either on unenhanced or gadolinium chelates—enhanced MR images. Furthermore, four central scars and four pseudocapsules were shown and led to a more confident diagnosis.
The clinical application of these new MR contrast agents must evolve into an integrated diagnostic scheme as a problem-solving tool in patients with atypical features of focal nodular hyperplasia and as a supplement to information provided by the use of nonspecific extracellular gadolinium chelates (e.g., differentiation of other types of hypervascular tumors). Vast clinical experience has been guided 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 lesions have rendered gadolinium chelates the first choice and most commonly used agents. The use of liver-specific contrast agents can help identify many morphologic and functional features of focal nodular hyperplasia lesions. For instance, homogeneous lesion enhancement after injection of a liver-specific contrast agent does not distinguish between benign and well-differentiated malignant hepatocellular tumors but allows better detection of morphologic patterns, such as a central scar in a lesion or homogeneity. In our series, the combined use of nonspecific and liver-specific contrast agents increased the confident diagnosis of lesions with atypical features, such as lesions that appeared hyperintense on the T1-weighted MR images or did not show typical uptake of gadolinium chelates during the dynamic arterial phase.
The definitive diagnosis of focal nodular hyperplasia can be made if lesions appear nearly isointense on unenhanced T1-weighted and T2-weighted images, if a central scar is hyperintense on T2-weighted images, and if lesions enhance early and vigorously after gadolinium chelates administration while enhancement of the central scar is delayed. However, in lesions without these pathognomonic appearances, administration of liver-specific MR contrast agents (either hepatobiliary or reticuloendothelial) is recommended. Homogeneous uptake of liver-specific MR contrast agents and better delineation of the central scar are useful in corroborating the diagnosis in cases with equivocal findings (Fig. 5).
|
Our study has several limitations including a lack of histologic proof of every lesion in every patient. In patients with multiple focal nodular hyperplasia lesions with the same appearance, only one lesion was proven by biopsy. In some patients with a single lesion, the diagnosis was based on imaging findings of correlative studies, constant size at subsequent examinations, or both. The lack of histologic confirmation of all focal nodular hyperplasia tumors can be explained by the fact that performing invasive procedures in patients with lesions that exhibit imaging features indicative of a benign histology cannot be justified.
The exact number and nature of the hepatic tumors could not be determined because none of the patients underwent liver transplantation, and none underwent autopsy. The total number of lesions detected on MR images obtained with each unenhanced and enhanced MR sequence at consensus interpretation was used to judge the detection and characterization of the hepatic focal nodular hyperplasia lesions. Because our purpose was to evaluate the MR imaging features of focal nodular hyperplasia in these patients and not to evaluate the accuracy of the diagnostic modality or the observers, we did not test for interobserver disagreement but used consensus opinion. Both observers agreed on the number of lesions and morphologic features. However, within the limits of ethical patient care, we believe that we have compelling evidence of the benign nature and likely histologic characteristics in all our cases.
Another limitation is that our series included only focal nodular hyperplasia lesions that were at least 1.5 cm. Reliable characterization of smaller lesions can be difficult, even with MR imaging [29]. Moreover, it has been shown that small liver lesions (<1.5 cm) are rarely malignant in the absence of a history of underlying malignancy or hepatic disease [30].
Our study was a retrospective descriptive study meant to establish reliable MR criteria for identifying focal nodular hyperplasia using different MR contrast agents. We believe that we now have the basis to conduct a prospective trial to test the accuracy of combined liver-specific and nonspecific contrast agents for MR imaging in this setting. Another study using multiple contrast agents in a single setting to diagnose questionable liver lesions should be performed. One report of this approach showed no clinically important negative interaction between the two contrast agents studied [31] and that this technique had cost—benefit and clinical significance in such problem cases.
In conclusion, contrast-enhanced MR images can show more functional and morphologic features of focal nodular hyperplasia than unenhanced MR images, thus allowing more confident diagnosis. In atypical lesions, the combined use of liver-specific and non-specific contrast agents may aid in noninvasive diagnosis of focal nodular hyperplasia.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
15 mm) hepatic lesions detected by CT.
AJR
1992;158:535
-539
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ferlicot, H. Kobeiter, J. T. Van Nhieu, D. Cherqui, D. Dhumeaux, D. Mathieu, and E. S. Zafrani MRI of Atypical Focal Nodular Hyperplasia of the Liver: Radiology-Pathology Correlation Am. J. Roentgenol., May 1, 2004; 182(5): 1227 - 1231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. D. Levy Invited Commentary RadioGraphics, January 1, 2004; 24(1): 18 - 19. [Full Text] [PDF]
|
|
| ||||||||||||||
