| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of North Carolina at Chapel Hill, CB 7510,
Chapel Hill, NC 27599-7510.
2 Department of Public Health, University of North Carolina at Chapel Hill,
Chapel Hill, NC 27599-7510.
3 Department of Pathology, University of North Carolina at Chapel Hill, Chapel
Hill, NC 27599-7510.
Received November 15, 2002;
accepted after revision March 4, 2003.
Address correspondence to R. C. Semelka.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Over a period of 52 months, we used our clinical information system to retrospectively identify the first MRIs obtained in 165 consecutive patients who had untreated liver metastases. All patients had histologic confirmation of the primary tumor. Liver metastases were confirmed at histologic examination, on imaging, or at clinical follow-up. MR sequences used included T1-weighted spoiled gradient-echo, T2-weighted half-Fourier acquisition single-shot turbo spin-echo, and serial gadolinium-enhanced spoiled gradient-echo imaging. Size, signal intensity characteristics, and pattern of enhancement of the metastases on MRIs were evaluated by two radiologists in consensus. Lesions were categorized by size: smaller than 1.5 cm, between 1.5 and 3.0 cm, and larger than 3.0 cm.
RESULTS. A total of 516 metastases (size range, 5-120 mm; mean, 28 mm) were assessed. Fifty-nine patients had hypervascular lesions, and 106 patients had hypovascular lesions. A significant difference in proportion of tumor vascularity was observed between the primary tumors described as classically hypervascular and those described as classically hypovascular (chi-square test for proportions of 70.8, p < 0.0001). The most common pattern was peripheral ring (72% of patients) seen on the arterial dominant phase images, with incomplete central progression (63%) seen on the delayed phase images. A hypointense ring seen in the periphery of the tumor during the delayed phase was the most common appearance in hypervascular metastases (27% patients) and was particularly conspicuous in patients with neuroendocrine and carcinoid tumors. Perilesional enhancement was common (47%), mostly seen in hypovascular metastases (92%). Generally, large lesions tended to show a peripheral ring or heterogeneous enhancement, and small lesions showed homogeneous enhancement.
CONCLUSION. MRI allows the identification of a wide spectrum of appearances of untreated liver metastases. The extent and pattern of enhancement of various histologic types of tumor are depicted on MRI.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Differences in the degree of vascularity among various types of hepatic lesions are the foundation for the enhancement-based recognition pattern and differentiation between benign and malignant lesions. Iodinated contrast agents and gadolinium chelates are extracellular agents that reflect the kinetics of tumor vascularization such as perfusion and diffusion through the capillaries into the extracellular space. The differences in the degree of enhancement thus reflect differences in the number of vessels, in the permeability of the vessels, and in the size of the extracellular space [3]. Other characteristics such as the presence of fibrous tissue or a high cellular tumoral density contribute to the pattern of enhancement [4].
From the radiologist's point of view, metastases represent a heterogeneous group of lesions that show variable signal characteristics and patterns of enhancement on early and late gadolinium-enhanced images reflecting the histopathologic nature of the tumor and its vascular component [1, 4]. However, some enhancement characteristics, such as an early ring enhancement and peripheral washout, are considered specific to metastases and allow differentiation of metastases from benign liver lesions such as cysts or hemangiomas [5, 6]. Other enhancement characteristics of liver metastases, such as perilesional enhancement, are less well recognized [7]. The excellent contrast and temporal resolution of MRI allow us to detect and characterize liver lesions with high accuracy and to differentiate benign from malignant lesions. Furthermore, certain metastases are generally considered either hypovascular (e.g., colon carcinoma) or hypervascular (e.g., breast carcinoma). To our knowledge, the percentage of metastases that fit into this vascularity categorization on the basis of their underlying histology has not been previously ascertained. The purpose of our study was to describe the spectrum of MRI appearances of liver metastases from different primary neoplasms, determining their degree of vascularity, the frequency of classically described enhancement patterns, and the impact of tumor size on the pattern of enhancement.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All patients had a histologic diagnosis of malignant disease that was based on examination of specimens from either the primary site or the liver. In 84 patients, the final diagnosis of liver lesions was established by biopsy of at least one lesion and in 27 patients, by partial hepatectomy for tumor resection. Other lesions with imaging findings similar to those of the histologically examined lesions were considered to be the same disease. In 54 patients without histologic evidence of liver tumors, an increase in the size or number of lesions was confirmed on follow-up imaging studies—MRI for 43 patients and CT for 11 patients—during a period of between 3 and 12 months.
The primary tumors were colon carcinoma (n = 66), pancreatic adenocarcinoma (n = 21), carcinoid tumor (n = 16), breast carcinoma (n = 16), pulmonary carcinoma (small cell type [n = 3], non-small cell type [n = 8]), renal cell carcinoma (n = 5), pancreatic neuroendocrine tumor (n = 6), adenocarcinoma of the small bowel (n = 2), gastric adenocarcinoma (n = 2), thyroid carcinoma (n = 2), prostate carcinoma (n = 2), transitional cell carcinoma of the bladder (n = 3), and unknown primary tumor (n = 13).
MRI Protocol
MRI of the upper abdomen was performed using a 1.5-T scanner (Vision,
Siemens Medical Solutions, Iselin, NJ) with a phased array multicoil for the
body, section thickness of 7-10 mm with a 20% intersectional gap, matrix size
of 128-192 x 256 (phase frequency encoding) for all sequences. All MR
examinations were performed using a set protocol that included breath-hold
T1-weighted spoiled gradient-echo (TR range/TE range, 120-170/4.0-4.5; flip
angle, 80-90°), breath-hold T2-weighted half-Fourier acquisition
single-shot turbo spin-echo (HASTE) (TR/TE, infinite/90; 2-3 acquisitions),
and breath-hold gadolinium-enhanced multiphasic T1-weighted spoiled
gradient-echo (120-170/4.0-4.5; flip angle, 80-90°) sequences. Using a
power injector (Medrad, Pittsburgh, PA), we injected 0.1 mmol/kg of gadolinium
chelate (Magnevist [gadopentetate dimeglumine], Berlex, Wayne, NJ; or Omniscan
[gadodiamide], Nycomed, Princeton, NJ) as a rapid bolus at 2 mL/sec. Serial
spoiled gradient-echo images were acquired at 18 sec (hepatic arterial
dominant phase) and at 45-60 sec (portal venous phase). Fat-suppressed spoiled
gradient-echo MRIs were acquired at 90-120 sec after contrast administration
(delayed phase) in all patients.
Image Review
Two radiologists retrospectively reviewed the MRIs to determine the size
and signal characteristics of the metastases on unenhanced T1-weighted and
T2-weighted images and the enhancement pattern on serial contrast-enhanced
images. The two radiologists were aware of the specific type of primary tumor
for each of the patients but were not provided with any other information
about the patients' histories. After each radiologist independently reviewed
the MRIs, agreement on the imaging findings was reached by consensus. Lesions
exhibiting pathognomonic findings of a cavernous hemangioma or cyst were
ignored during the MRI review.
All metastatic lesions were grouped into three categories according to the size of the diameter: smaller than 1.5 cm, between 1.5 and 3.0 cm, and larger than 3.0 cm. In patients in whom metastases were numerous, only two representative lesions were analyzed in each category. On T1-weighted images, the signal intensity of a lesion was categorized as mildly hypointense if it was between the level of intensity of the liver and that of the spleen, moderately hypointense if the intensity was comparable to that of the spleen, and markedly hypointense if the intensity was comparable to that of the cerebrospinal fluid. On T2-weighted images, the signal intensity of a lesion was categorized as mildly hyperintense if it was between the level of intensity of the liver and that of the spleen, moderately hyperintense if it was comparable to that of the spleen, and markedly hyperintense if it was comparable to that of the cerebrospinal fluid.
The three phases of enhancement were based on the location of gadolinium enhancement within the various hepatic vessels. The arterial dominant phase was defined as the point at which the contrast material was present in the arteries and in the main portal vein but not in the hepatic veins. The portal venous phase was defined as the point at which contrast material was in the arteries as well as in the portal and hepatic veins. The delayed phase was defined as occuring 2-3 min after contrast administration. Contrast enhancement of the lesions was evaluated on the basis of the pattern of uptake of contrast material during the arterial dominant phase and the presence of centripetal progression of enhancement and washout during the portal venous and delayed phases.
In the arterial dominant phase, a lesion was defined as hypervascular if its enhancement was greater than that of liver parenchyma and equal to that of pancreatic parenchyma and as hypovascular if its enhancement was less than or equal to that of the liver parenchyma. The pattern of enhancement was categorized as diffusely homogeneous, heterogeneous, peripheral ring, or negligible. If the size of a lesion on unenhanced images was essentially identical to its size on images obtained immediately after contrast administration, the enhancement of the lesion was judged to have no contribution to perilesional enhancement.
Perilesional enhancement was defined as enhancement that was seen beyond the lesion margins in the surrounding liver parenchyma when images obtained immediately after contrast material administration were compared with unenhanced T1-weighted images. Perilesional enhancement was further defined as wedge-shaped or circumferential.
The progression of enhancement over time was determined during the portal venous and delayed phases; it was considered to be incomplete if the central portion of the tumor lacked enhancement and to be complete if the total volume of the tumor showed enhancement that was either isointense or hyperintense relative to liver parenchyma. Washout was considered to have occurred if the lesion showed a decrease in enhancement, with hypointensity in comparison to liver parenchyma during the delayed phase.
The Mantel-Haenszel chi-square test was performed to compare differences in the proportions of hypervascular liver metastases in the tumor population classically considered to be hypovascular (colon adenocarcinoma, bladder adenocarcinoma, prostate adenocarcinoma, and pulmonary carcinoma) and in the proportions of hypovascular liver metastases in the tumor population classically considered to be hypervascular (thyroid carcinoma, carcinoid tumor, neuroendocrine tumors, and renal cell carcinoma).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The enhancement patterns associated with each type of metastases are summarized in Table 1. A hypervascular pattern of enhancement was identified in liver metastases in 59 (36%) of the 165 patients, including six (9%) of the 66 patients with colon adenocarcinoma (Figs. 1A, 1B, 1C, and 1D), one (50%) of the two patients with gastric adenocarcinoma, seven (33%) of the 21 patients with pancreatic adenocarcinoma, 11 (69%) of the 16 patients with breast carcinoma (Figs. 2A, 2B, 2C, and 2D), 14 (89%) of the 16 patients with carcinoid tumor (Figs. 3A, 3B, 3C, and 3D), all six patients (100%) with neuroendocrine tumors, four (80%) of the five patients with renal cell carcinoma, both (100%) of two patients with thyroid carcinoma, and eight (62%) of the 13 patients with an unknown type of primary tumor.
|
|
|
|
|
|
|
|
|
|
|
|
|
A hypovascular pattern of enhancement was identified in liver metastases in 106 (64%) of the 165 patients, including 60 (91%) of the 66 patients with colon adenocarcinoma (Figs. 4A, 4B, 4C, 4D, and 4E), one (50%) of the two patients with gastric adenocarcinoma, both (100%) of the two patients with small-bowel adenocarcinoma, 14 (67%) of the 21 patients with pancreatic adenocarcinoma, five (31%) of the 16 patients with breast carcinoma, two (11%) of the 16 patients with carcinoid tumor, all 11 patients (100%) with pulmonary carcinoma, one (20%) of the five patients with renal cell carcinoma, all three patients (100%) with bladder carcinoma, both (100%) of the two patients with prostate carcinoma, and five (28%) of the 13 patients with an unknown type of primary tumor.
|
|
|
|
|
The two most common patterns of enhancement identified on the arterial dominant phase in both hypervascular and hypovascular metastases were peripheral ring enhancement (119/165 or 72%) and heterogeneous enhancement (28/165 or 17%). Heterogeneous enhancement was seen in lesions with diameters larger than 3.0 cm; these included hypovascular metastases in six patients with colon carcinoma, hypervascular metastases in one patient with thyroid carcinoma, and two patients with renal cell carcinoma. A homogeneous pattern was identified in 18 (39%) of 59 patients with hypervascular lesions smaller than 1.5 cm (Figs. 3A, 3B, 3C, and 3D). Typically, large lesions showed peripheral ring or heterogeneous enhancement, and small lesions showed homogeneous enhancement. Negligible enhancement was identified in the lesions of 19 (32%) of 106 patients having hypovascular metastases. For all hypervascular metastases, the degree of enhancement was similar to that of the pancreas on the arterial dominant phase MRIs. No internal vessels or discontinuous peripheral nodular pattern of enhancement was identified.
On the portal venous and delayed phase MRIs, incomplete central progression of lesion enhancement was found in 105 (64%) of the 165 patients. Progression to isointensity was identified in 21 patients (13%). Progression to central hyperintensity was identified in 34 patients (21%). Persistent negligible enhancement was identified in five patients (3%). Peripheral washout was identified on images obtained during the late phases of enhancement in 18 patients—16 (27%) of the 59 patients with hypervascular metastases (Figs. 2A, 2B, 2C, 2D and 3A, 3B, 3C, 3D) and two (2%) of the 106 patients with hypovascular metastases. Perilesional enhancement was identified in 79 patients (48%) and was associated with hypovascular liver metastases in 70 patients, with metastases from colon carcinoma in 39 patients, and with pancreatic adenocarcinoma in 19 patients. A wedge-shaped pattern was present in 41 patients and a circumferential pattern was identified in 38 patients.
Some particular characteristics were identified in metastases from colon adenocarcinoma. All hypovascular lesions, regardless of size, had a thin peripheral ring of strong enhancement that persisted through all phases of enhancement. Histopathologic correlation obtained in 13 patients showed a thin zone containing varying amounts of fibrous and inflammatory cells. All colon cancer metastases with diameters larger than 3 cm had a cauliflower appearance on gadolinium-enhanced MRIs and strands of fibrous tissue and inflammatory cells that extended into the lesion surrounding islets of tumoral cells (Figs. 4A, 4B, 4C, 4D, and 4E).
Among the liver metastases from primary tumors classically considered hypovascular, we found six hypervascular liver metastases (7%) and 76 hypovascular liver metastases (93%). Among the liver metastases from primary tumors classically considered hypervascular, we found 26 hypervascular liver metastases (90%) and three hypovascular metastases (10%). We noted a significant difference in the proportion of tumor vascularity for metastases between the primary tumors classically described as hypervascular and those classically described as hypovascular (chi-square test for proportions of 70.8, p < 0.0001).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although the concepts that liver metastases have an arterial supply and that the degree of enhancement on radiologic images obtained during the arterial dominant phase is mostly related to the degree of vascularization and perfusion of the lesion are generally accepted [2], comparing the degree of lesion enhancement with that of the liver may be problematic. Enhancement in liver parenchyma varies depending on the timing of contrast material injection or variants in the vascular supply of the liver. Another way to assess the vascularity of a hepatic tumor is to compare the degree of enhancement of the lesion with that of a highly vascularized organ. We selected the pancreas as the standard of reference with which to assess the vascularity of liver metastases. Because the normal pancreas is a richly vascularized organ that truly reflects capillary flow during the arterial phase, it may be considered a reasonable indicator of the degree of enhancement of a lesion and thus of its perfusion. An additional benefit of using the pancreas as a reference is that it is located in the same tissue volume as the liver on axial MRIs, so direct pancreas-to-lesion comparisons can be made. Using reference organs that are in a different position along the slice selection may introduce errors in comparing levels of enhancement because of variations in the intrinsic signal-to-noise ratios. In our study, all hypervascular lesions showed enhancement comparable to that of the pancreas on the arterial dominant phase images. We suggest that in future studies of liver tumor vascularity, the pancreas be used as a reference organ for comparison of liver lesion enhancement during the arterial dominant phase.
All our patients had mildly hypointensive lesions on T1-weighted images; additionally, 15 patients had isointense lesions. On T2-weighted images, 72% of patients had metastases that were moderately hyperintensive, and 20% of patients had metastases with areas of marked hyperintensity. Generally, metastases have a variable morphologic appearance and signal intensity on T1- and T2-weighted images [4]. On T2-weighted images, several morphologic patterns have been described for liver metastases [8]. Histopathologic correlation with the signal intensity seen on T2-weighted images showed that areas of marked hyperintensity corresponded to cystic changes or necrosis [9]. In our study, a specific morphologic pattern on T2-weighted images was not related to metastases from a specific primary neoplasm. We found much overlap among the appearances of metastases from different primary origins.
Generally, the vascularity of the metastases has been considered to be similar to the vascularity of the primary neoplastic lesion [10, 11]. Classically, metastases from colon carcinoma, bladder carcinoma, prostate carcinoma, and pulmonary carcinoma have been considered hypovascular, and those from thyroid carcinoma, carcinoid tumor, neuroendocrine tumor, and renal cell carcinoma have been considered hypervascular [4, 11]. We found a strong association between the classically described vascularization of the primary tumor and the vascularization of the liver metastases. In the literature, no clear consensus emerges as to the vascularity of liver metastases from breast carcinoma. Breast carcinoma metastases are described as hypervascular by some authors [11, 12], whereas others found hypovascular metastases to be more common [13]. In our study, 69% of patients with metastases from breast carcinoma showed a hypervascular pattern during the arterial dominant phase. Although metastases from pancreatic adenocarcinoma are traditionally considered hypovascular, hypervascular metastases were identified in 33% of the patients in our study. This result may, in part, be related to the presence of a hypervascular primary tumor, as is the case with up to 45% of pancreatic adenocarcinomas reported in the pathology literature [14, 15]. However, a complex series of interactions between the tumor and the host microenvironment have been described that influence the degree of vascularization of primary tumors and metastases, and we believe these interactions may be the major reason for the difference in vascularization between the primary tumor and the metastatic lesions in patients with breast carcinoma and pancreatic adenocarcinoma [14-16].
An accurate differentiation between benign liver lesions and metastases is possible using specific enhancement patterns [4]. Certain patterns of enhancement allow differentiation of benign liver lesions with high specificity [5, 6]. In our study, no metastases showed the classic enhancement pattern of benign lesions such as the absence of enhancement in simple cysts or the peripheral progressive discontinuous nodular enhancement that is highly specific for hemangiomas. Four major metastatic enhancement patterns in the arterial dominant phase were seen in our patients: peripheral ring, homogeneous, heterogeneous, and negligible enhancement. In our study, a peripheral ring was the most common pattern of enhancement on the arterial dominant phase MRIs in both hypervascular and hypovascular metastases, identified in 72% of our patients. MRI is highly sensitive to gadolinium; therefore, most of the hypovascular metastases showed enhancement. The peripheral ring pattern is considered a specific enhancement pattern for liver metastases [5, 6]. The peripheral ring enhancement may reflect the pattern of growth of metastases with a parasitic blood supply taken from the surrounding liver, regardless of the abundance of vascular supply. As the tumor reaches a threshold size, only the peripheral part remains well vascularized, whereas cells in the central portion of the tumor lose the close proximity with the vascular supply and become compressed, necrotic, or replaced by fibrosis.
Hypervascular metastases smaller than 1.5 cm showed homogeneous early enhancement, and the differentiation of these metastases from homogeneously enhancing hemangiomas smaller than 1.5 cm may be problematic, especially in the absence of other larger metastases [17]. Homogeneous early enhancement was observed in several patients (39%) with hypervascular metastases, a finding that emphasizes the importance for close imaging follow-up of small lesions in patients with a known history of malignancy. Small lesions attributable to cavernous hemangioma, adenoma, focal nodular hyperplasia, or arterioportal shunting also can show homogeneous early enhancement, and biopsy of such lesions is often difficult. We observed a heterogeneous pattern of enhancement in 17% of patients, mostly in those with lesions with diameters larger than 3 cm. The heterogeneous pattern is not specific for metastases; the pattern can be observed in primary malignant liver lesions, especially hepatocellular carcinomas [4]. Negligible enhancement is also common, identified in 32% of our patients with hypovascular metastases.
Incomplete central progression was the most common enhancement pattern on delayed phase images in both hypervascular and hypovascular metastases, found in 63% of our patients. The incomplete central progression may be related to a slow blood flow within the central part of the tumor, which, as discussed earlier, is poorly vascularized. Progression to isointensity on delayed phase images was less common (13%), found in both hypervascular and hypovascular metastases, and may be related to the passage of the contrast agent into larger interstitial spaces.
Peripheral washout in metastases on delayed phase images has been described as a specific sign for metastases (specificity, 100%) [18]. Although present in both hypervascular and hypovascular lesions, peripheral washout was more conspicuous and observed more frequently in hypervascular metastases, particularly in neuroendocrine and carcinoid tumors. Peripheral washout may reflect a transitory perfusion within the well-vascularized peripheral portion of the lesion, which has both good arterial supply and good venous withdrawal of contrast material.
We found that, with the exception of metastases from colorectal carcinoma, the enhancement patterns in metastases from different primary neoplasms overlap, with the most common pattern being that of peripheral ring on the arterial dominant phase MRIs and incomplete centripetal progression on the delayed phase MRIs. The cauliflower appearance with scalloped margins and enhancement of internal septations have been described as characteristic of metastases from colon carcinoma [4]. In our study, this appearance was found in all colon carcinoma metastases larger than 3 cm and was correlated with the presence of fibrotic and inflammatory strands extending into the tumor and surrounding tumoral islets. Regardless of the size of the lesion, colon carcinoma metastases on the arterial dominant phase MRIs showed a distinctive marginal thin ring of strong enhancement that persisted through the delayed phase of enhancement. This feature was observed only in colon carcinoma metastases, and in our previous study [7], histopathologic evaluation showed that this ring represents a thin margin of a tumoral zone composed of varying amounts of fibrous tissue and inflammatory cells. As the tumor grows, the inflammatory and fibrous septa extend within the central part of the tumor, giving the lesion its classic cauliflower appearance.
Transient perilesional enhancement was a common finding in our study, identified in 47% of the patients. Such enhancement was most conspicuous on MRIs obtained during the arterial dominant phase. This enhancement has been shown to correlate with histopathologic hepatic parenchymal changes, including peritumoral desmoplastic reaction, inflammatory cells, and vascular proliferation [7]. The peritumoral enhancement can be wedge-shaped or circumferential. The enhancement was most commonly identified in liver metastases from pancreatic and colorectal adenocarcinoma. The wedge-shaped pattern of enhancement results from an increase in the regional arterial flow. However, the precise underlying pathophysiologic cause remains uncertain. We believe that it is related to a vascular imbalance between the arterial and the portal blood supplies, with a resultant increase of the hepatic arterial supply to tumors or portal venous obstruction proximal to tumors because of compression or tumor invasion [19]. In our study, perilesional enhancement was less commonly seen in hypervascular lesions (8%), suggesting that an increase in the hepatic arterial supply alone may not be the cause of the perilesional enhancement.
Our study has some limitations. Histologic proof was not obtained for all lesions. We had only a few patients with liver metastases from some of the primary neoplasms such as thyroid, bladder, or prostate carcinomas. In addition, over the period of time studied, no patients with melanoma were imaged; melanoma can have a distinctive hyperintensity on T1-weighted images. Other histologic types of liver metastases that may appear hyperintense on T1-weighted images were not in our study population, reflecting the fact that we had a single-institution case accrual and that only patients with untreated metastases were included. On the basis of our initial observations, we believe that conducting a similar study in a larger population would be interesting.
A crucial part of our study protocol was to include only patients with untreated liver metastases. At our institution, most of the patients who undergo MRI for liver metastases have already been treated with chemotherapy, and some have received local ablative therapy. Therefore, applying this criterion greatly limited the number of patients in our study, but we considered this element of the study to be essential because chemotherapy induces vascular changes in liver metastases that greatly alter the appearance of the lesions [20].
The MRIs we reviewed were obtained with a set 18-sec delay between initiation of contrast material injection and image acquisition. In consideration of patient throughput, no bolus-tracking device or test bolus imaging was used to optimize the timing for the arterial dominant phase. This practice may have affected the image quality in some patients. However, maximal image quality occurs when contrast material first appears in the portal veins, and all current timing schemes are based on hepatic artery enhancement, which occurs earlier than portal vein enhancement. It is not clear to us why our method would be inferior to others currently used.
Visual assessment of appropriateness of the phase of contrast enhancement was performed retrospectively. Some researchers suggest that the breath-hold turbo spin-echo sequence is not sensitive to the subtle difference in T2 contrast of soft tissue and that reliability of the sequence in revealing hepatic tumors is inadequate [21, 22]. We routinely use a breath-hold HASTE sequence with short acquisition time because we rely predominantly on T1-weighted images and early phase gadolinium-enhanced MRIs for lesion detection and characterization for consistent image quality and reasonable patient throughput. We understand that different metastases in the same liver can show different MRI findings. In patients with multiple metastases, we analyzed two lesions that were representative of the imaging findings of most of the lesions in the liver. In most patients, these two lesions were the largest in the liver. Furthermore, in patients for whom we had only two lesions, we did not observe any differences in imaging findings between these lesions.
Our study had limitations related to the retrospective study design and consensus interpretation. The two radiologists reviewing the MRIs were aware of the specific types of primary tumors at the time of the image review, which might have biased their findings to a certain extent. In addition, we first selected patients using the diagnostic code for liver metastasis to search the clinical information system in our hospital and then identified the patients with untreated metastases who underwent MRI. The selection bias was considered minimal because in our institution, most patients with focal hepatic disease undergo MRI of the liver.
MRI can display the spectrum of appearances of untreated liver metastases. Dynamic gadolinium-enhanced MRIs are particularly useful in showing the different patterns of enhancement. With the exception of metastases from colon carcinoma with the characteristic cauliflower appearance on gadolinium-enhanced MRIs, we observed considerable overlap among the appearances of metastases from different primary origins. The most common pattern of enhancement of liver metastases was that of a peripheral ring on the arterial dominant phase images, with incomplete central progression on the delayed phase images. Hypervascular and hypovascular lesions share most of the enhancement patterns, reflecting the fact that the growth pattern of the tumor and the derivation and pattern of blood supply are not dependent on the degree of vascularization. The concordance between the vascularity of the liver metastases and the vascularity classically described for the primary tumor of origin was statistically significant. However, patients with metastases considered either classically hypovascular or classically hypervascular may have liver metastases that exhibit the opposite type of enhancement; for example, 9% of patients with colon carcinoma had hypervascular metastases. Peripheral washout is common in hypervascular metastases and is particularly conspicuous in neuroendocrine and carcinoid tumors. Transitory perilesional enhancement—either wedge-shaped or circumferential—was frequently observed and was most commonly found in hypovascular metastases.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Parikh, S. J. Drew, V. S. Lee, S. Wong, E. M. Hecht, J. S. Babb, and B. Taouli Focal Liver Lesion Detection and Characterization with Diffusion-weighted MR Imaging: Comparison with Standard Breath-hold T2-weighted Imaging Radiology, March 1, 2008; 246(3): 812 - 822. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Hughes-Cassidy, J. Wong, D. Aguirre, A. D. Chavez, T. Wolfson, A. Gamst, and C. Sirlin Transient Homogeneously Enhancing Hepatic Masses: Can Size Predict Benignity? Am. J. Roentgenol., February 1, 2008; 190(2): 300 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Giraudet, D. Vanel, S. Leboulleux, A. Auperin, C. Dromain, L. Chami, N. Ny Tovo, J. Lumbroso, N. Lassau, G. Bonniaud, et al. Imaging Medullary Thyroid Carcinoma with Persistent Elevated Calcitonin Levels J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4185 - 4190. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kanematsu, S. Goshima, H. Kondo, Y. Tsuge, R. Yokoyama, K. Kajita, M. Onozuka, Y. Suzuki, M. Van Cauteren, and N. Moriyama Gadolinium-Enhanced Multiphasic 3D MRI of the Liver with Prospective Adaptive Navigator Correction: Phantom Study and Preliminary Clinical Evaluation Am. J. Roentgenol., April 1, 2007; 188(4): W309 - W316. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. Elsayes, V. R. Narra, Y. Yin, G. Mukundan, M. Lammle, and J. J. Brown Focal Hepatic Lesions: Diagnostic Value of Enhancement Pattern Approach with Contrast-enhanced 3D Gradient-Echo MR Imaging RadioGraphics, September 1, 2005; 25(5): 1299 - 1320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Grazioli, G. Morana, M. A. Kirchin, and G. Schneider Accurate Differentiation of Focal Nodular Hyperplasia from Hepatic Adenoma at Gadobenate Dimeglumine-enhanced MR Imaging: Prospective Study Radiology, July 1, 2005; 236(1): 166 - 177. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
