| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Beth Israel Deaconess Medical Center, Harvard Medical
School, 330 Brookline Ave., Boston, MA 02215.
2 Present address: Department of Diagnostic Radiology, Yonsei University College
of Medicine, Yongdong Severance Hospital, Seoul, Republic of Korea
135-720.
Received November 1, 2004;
accepted after revision February 9, 2005.
Address correspondence to: J-S Yu
(yjsrad97{at}yumc.yonsei.ac.kr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. One hundred thirty-four lesions of hepatic metastases in 44 patients were subjected to a retrospective analysis of the dynamic MR images obtained with 3D spoiled gradient-echo sequences. The thickness of the enhancing rim on arterial phase images was regarded as a summation of the enhancing component of tumor periphery and perilesional enhancement, which were estimated by the tumor size on precontrast T1-weighted images. The presence of wedge-shaped perilesional enhancement was also correlated with the lesion size.
RESULTS. Except for 17 diffusely enhanced lesions, lesion size was comparable between the lesions with (n = 87; 26 ± 19 [SD] mm) and without rim enhancement (n = 30; 27 ± 23 mm) on the arterial phase dynamic MR images (p > 0.05). The degree of peripheral tumoral enhancement showed an inverse correlation (r = -0.389) with the thickness of the circumferential perilesional enhancement (p < 0.001). The mean size of the lesions with wedge-shaped perilesional enhancement (n = 44; 33 ± 20 mm) was larger than that of the other lesions (n = 90; 25 ± 19 mm) (p = 0.016).
CONCLUSION. The degree of circumferential perilesional enhancement of hepatic metastases on arterial phase dynamic MR images would be independent of the lesion size but inversely correlated with the degree of peripheral tumoral vascularity. An understanding of these features may help tumor characterization and should prompt hypotheses and studies of microvascular phenomena in tumoral and epitumoral environments.
Keywords: abdominal imaging • dynamic MRI • liver • liver disease • MRI
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To our knowledge, the incidence and pattern of perilesional enhancement in arterial phase dynamic MR images have yet to be determined in a large group of patients with hepatic metastases. The purpose of this study was to determine whether perilesional parenchymal enhancement of hepatic metastases was correlated with the degree of tumoral enhancement on arterial phase images or tumor size using dynamic MRI. In addition, the insights gained by our observations are discussed in the context of results from previous reports, aimed toward a further understanding of the perilesional enhancement of hepatic metastases.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The criteria for entry into the analysis were identification of discrete focal lesions, successful arterial phase MRI that revealed contrast filling in intrahepatic portal veins without hepatic vein enhancement and excluding technical failure, no previous localized percutaneous ablation therapy for focal lesions, and clearly visible lesions on unenhanced MR images to facilitate sizing independent of contrast media effects. For case selection, digitally stored image data in the PACS were retrieved and preliminarily reviewed by one radiologist in conjunction with the medical records of the 56 patients. In four patients, arterial phase imaging was not available because of a technical failure or poor image quality resulting from motion artifacts during the imaging. Diffusely infiltrating lesions affecting the entire liver prevented lesion-by-lesion analysis in three patients. Another four patients had a previous history of arterial chemoembolization (n = 1), percutaneous ethanol injection (n = 1), or radiofrequency ablation (n = 2) for all visible lesions. The isosignal intensity of the lesions on precontrast T1-weighted images prevented the size estimation for one other patient. Thus, 12 patients were excluded from this study based on a failure to meet the entry criteria.
Of the remaining 44 patients (26 men and 18 women; age range, 34–87 years; mean, 62 years), the hepatic lesions and primary extrahepatic lesions were histologically diagnosed in 21 patients. For the other 23 patients, the primary lesions were pathologically verified and the hepatic lesions showed definite interval growth on serial imaging studies. The primary cancer in the 44-patient cohort with hepatic metastases was 19 colorectal cancers, nine pancreatic cancers, seven breast cancers, two lung cancers, one extrahepatic bile duct cancer, one urinary bladder cancer, one rectal carcinoid tumor, one esophageal cancer, one islet cell tumor, one melanoma, and one renal cell carcinoma.
Thirteen patients had single lesions, and 31 had multiple lesions. In patients with more than five lesions, the five most conspicuous lesions on unenhanced T1-weighted MR images were selected subjectively by one radiologist and used for analysis without previous knowledge of the findings on the arterial phase contrast-enhanced images of each lesion. A total of 134 lesions (colon cancer, n = 57; pancreatic cancer, n = 22; breast cancer, n = 22; lung cancer, n = 6; extrahepatic bile duct cancer, n = 5; esophageal cancer, n = 5; melanoma, n = 5; islet cell tumor, n = 5; carcinoid, n = 4; bladder cancer, n = 2; renal cell carcinoma, n = 1) were analyzed for perilesional enhancement on arterial phase dynamic MRI.
MRI
MRI was performed with a 1.5-T unit (Symphony or Vision, Siemens Medical
Solutions) with a torso phased-array coil. A 22-gauge IV catheter was placed
in an arm vein and attached to an MRI-compatible power injector (Spectris,
Medrad). T2-weighted imaging with the STIR turbo spin-echo technique (TR/TE,
3,500–4,000; inversion time, 65–80/165 msec; refocusing pulse,
130°; bandwidth, 325 Hz/pixel) was performed on the axial plane. After
phase-contrast imaging with a double echo spoiled gradient-echo sequence
(192/2.7 for opposed phase and 5.3 msec for in phase; flip angle, 80°;
bandwidth, 488 Hz/pixel), precontrast T1-weighted imaging and multiphase
dynamic contrast-enhanced dynamic imaging were performed with 3D spoiled
gradient-echo sequences.
The volumetric interpolated breath-hold examination (VIBE) [17], a 3D spoiled gradient-echo sequence (3.8–5.2/1.6–1.9; flip angle, 12°; bandwidth, 488 Hz/pixel) was used for dynamic imaging. The field of view was 300–370 mm with a rectangular configuration in the phase-encoding (anteroposterior) dimension. The imaging matrix was 256 x 256 with a pixel size of less than 1.5 mm. The slab thickness ranged from 160 to 200 mm to ensure full coverage of the liver. The partition thickness after interpolation was 1.4–3.2 mm with 72–128 partitions, and the acquisition time ranged from 18 to 28 sec, depending on the liver coverage, partition thickness, and the patient's breath-holding capacity. The VIBE incorporates a frequency-selected fat-saturation pulse before each partition loop. After obtaining a precontrast VIBE, a timing examination was performed according to a previously described method [18] by using a 1-mL test dose of gadopentetate dimeglumine (Magnevist, Berlex). For arterial phase imaging, all patients received a 19-mL bolus of contrast material. After the arterial phase, the VIBE was then repeated twice at 45-sec intervals after initiation of the preceding acquisition for portal vein and equilibrium phase imaging.
Image Analysis
Tumor size was estimated by using the longest dimension on precontrast VIBE
images. One experienced radiologist measured and recorded the tumor size on
magnified views of axial images using electronic calipers on the PACS monitor.
Perilesional enhancement was defined as circumferential or wedge-shaped high
signal intensity around or adjacent to the lesions distinguished from the
background hepatic parenchyma on contrast-enhanced images. The presence of rim
or wedge-shaped enhancement was determined through a comparison of precontrast
and arterial phase gadolinium-chelate-enhanced VIBE images performed by the
same radiologist who reviewed the entire imaging data for this selection of
patients.
After comparison and synchronization of the anatomic level of the arterial phase images with the precontrast images, the latter having already been used for measuring the tumor size, the outer and inner diameters were measured along the longest dimension for each lesion for which rim enhancement was present. The outer thickness of the rim enhancement was estimated by subtracting the unenhanced tumor size from the outer dimension of the rim enhancement, and the result was regarded as the thickness of circumferential perilesional enhancement (Fig. 1). Meanwhile, the inner thickness of the rim enhancement was estimated by subtracting the inner dimension of the rim enhancement from the unenhanced tumor size, and the result was used to represent the tumor vascularity in the periphery of each lesion (Fig. 1). For those lesions with diffuse enhancement and without any recognizable peripheral rim, perilesional enhancement was estimated by subtracting the tumor size on the unenhanced image from the size of the lesion on the arterial phase images; the inner dimension of the rim was considered zero in such lesions. To estimate the possibly isointense contrast enhancement of the periphery of the lesions without hyperintense rim or diffuse enhancement, only the lesional dimension of hypointense signal on the arterial phase image was subtracted from the lesion size on the corresponding precontrast image.
|
|
|
|
|
|
|
The Student's t test for unpaired data was performed to compare the sizes of the lesions accompanying the perilesional enhancement with the sizes of the lesions without perilesional enhancement, corresponding to the circumferential and wedge-shaped enhancements, respectively. The Student's t test for paired data was used for the comparison of the mean outer thickness, representing the circumferential perilesional enhancement (tumor size subtracted from outer diameter of the rim enhancement) with the inner thickness of the rim enhancement, representing the peripheral enhancing portion of the tumor itself (inner diameter of the rim enhancement subtracted from the tumor size) to determine the main component of rim enhancement. Pearson's correlation test was used to evaluate the correlation between the thickness of the rim enhancement and the tumor size. Pearson's correlation coefficient was also calculated to compare the thickness of the circumferential perilesional enhancement and the thickness of the peripheral enhancing portion of the tumor itself in each lesion. The one-way analysis of variance test was used to compare the tumor size among the three different types of perilesional wedge-shaped enhancement. For all tests, p < 0.05 was deemed to indicate statistical significance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Three of 17 lesions with diffuse enhancement of the entire lesion on the arterial phase dynamic MR images showed high signal intensities, enabling the measurement of their size. The other 14 lesions showed similar signal intensities poorly delineated from the surrounding hepatic parenchyma. The size of the hyperintensity on the arterial phase images of the three tumors was larger than that measured on the unenhanced T1-weighted images, which suggested the presence of perilesional enhancement (Table 1).
|
Thirty lesions without rim enhancement showed low signal intensities on arterial phase dynamic images. Among them, 10 lesions showed more than a 2-mm reduction in diameter on the arterial phase dynamic images, suggesting the presence of peripherally increased tumor vascularity with isointense contrast enhancement indistinguishable from the surrounding hepatic parenchyma. The other 20 lesions showed less than a 2-mm reduction in diameter (Table 1).
Except for 17 diffusely enhanced lesions on the arterial phase dynamic MR images, there was no significant size difference between lesions with hyperintense rim enhancement (n = 87; 26 ± 19 [SD] mm) and lesions without distinguishable rim enhancement (n = 30; 27 ± 23 mm) (p > 0.05).
For the 87 lesions with rim enhancement, the thickness of the outer portion of the rim enhancement (mean, 6 ± 3 mm) was significantly larger than that of the inner portion (mean, 1 ± 2 mm) (p < 0.001), which suggested that the main component of the rim enhancement was perilesional enhancement rather than tumoral enhancement. In these cases, the tumor size was not correlated with the thickness of the rim enhancement (p > 0.05). Meanwhile, the degree of tumoral enhancement (inner thickness of the rim enhancement) showed a significant inverse correlation (r = -0.389) with the thickness of the perilesional rim enhancement (p < 0.001). In other words, the thickness of the hypervascular portion of the tumor periphery was correlated with a thinner degree of perilesional rim enhancement compared with those tumors with a relatively thinner hypervascular periphery that were correlated with a thicker perilesional rim enhancement.
The mean size of all lesions accompanied by wedge-shaped perilesional enhancement (n = 44; 33 ± 20 mm) was larger than that of the other lesions (n = 90; 25 ± 19 mm) (p = 0.016). In regard to the extent of the perilesional enhancement, 26 lesions were categorized as the distal type (mean tumor size, 35 ± 22 mm), 11 lesions were proximal type (tumor size, 30 ± 18 mm), and another seven lesions were accompanied by nonsegmental type (tumor size, 27 ± 15 mm) wedge-shaped enhancements (Figs. 4A, 4B, and 4C). No correlation was seen between the lesion size and the type of wedge-shaped enhancement (p > 0.05).
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One interesting finding in the present study was that the circumferential perilesional enhancement was thicker in the lesions with thin peripheral tumoral enhancement and vice versa. In other words, the lesions with a higher vascular component showed a lesser degree of perilesional enhancement, and the lesions with lower vascularity showed a greater degree of perilesional enhancement. From the viewpoint of tumor characterization, this observation differs from the results of previous investigations involving other types of tumors, including hepatocellular carcinoma or hepatic cavernous hemangioma [20–22]. Ueda et al. [20] reported that perilesional corona enhancement was observed after complete filling of hypervascular hepatocellular carcinoma in their single level CT arteriography study. Moreover, in a report by Yu et al. [21] with dynamic MRI, perilesional enhancement tended to be found in larger and more hypervascular tumors in hepatocellular carcinomas. Comparing the high incidence of perilesional enhancement in this study of hepatic metastases (circumferential enhancement, 63%; wedge enhancement, 33%), the overall incidence of perilesional enhancement for uncomplicated hepatocellular carcinoma without gross portal vein invasion was considerably low, with a value of less than 12% [21]. For hepatic cavernous hemangiomas, Jeong et al. [22] reported that the hyperdynamic status of rapidly enhancing hemangiomas or intralesional nidus of arterioportal communication was regarded as the main cause of the perilesional enhancement in their MRI study. For the mechanism of rim enhancement of hepatic metastases, their theory of transtumoral shunt or tumor draining venous flow into the adjacent hepatic sinusoids [20–22] cannot be invoked to explain the cause of the perilesional circumferential enhancement of hepatic metastases.
In an animal microscopic study of metastasis model, Kan et al. [23] showed that tumor cells and inflammatory cells tended to adhere to the endothelial lining in the hepatic sinusoids at the tumor border, causing occlusion or stagnation of blood flow at the periphery of the tumor. Considering previous pathologic investigations in conjunction with the data of the present study, the circumferentially increased arterial flow around the hepatic metastases can be explained by a certain degree of increased physiologic or functional arterioportal shunts from the increased sinusoidal pressure due to compression [13], narrowing or obliteration of perilesional sinusoids by tumor cells, inflammatory cells [23], or desmoplastic reactions [11], in addition to a sinusoidal capillarization at the tumor border during the neoplastic angiogenesis [24]. In a recent report of single-level CT hepatic arteriography with pathologic correlation by Terayama et al. [12], hypovascular metastases tended to show early appearance of rim enhancement and hypervascular metastases showed more delayed rim enhancement. They proposed that the rim enhancement of hepatic metastases is caused by altered hemodynamics in the surrounding liver parenchyma for hypovascular lesions in addition to the drainage flow from the hypervascular lesions.
With regard to the reverse correlation between the thickness of the perilesional enhancement and the hypervascular area of the tumor, which was observed in this study, we can speculate that under conditions of limited arterial blood flow into a hepatic segment containing the tumor, a relatively larger portion of arterial blood can be introduced into the abundant tumor vessels in the metastatic lesion without resistance. During the enhancement of the tumor, a relatively smaller portion of the arterial blood is able to circulate into the tumor border containing the potential arterioportal shunt. In the case of metastases with a paucity of tumor vessels, however, the main proportion of the arterial blood can be supplied into the tumor border with an abundant arterioportal shunt, resulting in a thick perilesional enhancement. Single-level CT arteriography studies [12, 21] revealed that most hypervascular metastases and hepatocellular carcinomas showed more delayed peritumoral enhancement followed by substantial tumoral enhancement. These findings also implicate the low incidence and lesser degree of peritumoral enhancement on early arterial phase images for hypervascular tumors during conventional dynamic MRI.
In addition to characterizing the metastases from other hepatic tumors, these different features and mechanisms of perilesional rim enhancement can influence therapeutic planning. Although the extent of the perilesional enhancement is not matched with the extent of the substantial parenchymal change [11], the area of perilesional rim enhancement is open to the possibility of microscopic tumor cell infiltration, and we can justify the widening of safety margin for local ablation therapy or partial hepatic resection.
In the present study, the overall incidence of wedge-shaped parenchymal enhancement was lower than that of the circumferential enhancement and tended to be accompanied by larger lesions. As documented in a previous study [25], portal vein tributaries and hepatic vein outflow tracks are easily compressed by adjacent space-occupying lesions and can induce a compensatory increase in hepatic arterial flow [26] or a reversal of portal venous flow direction [27, 28], respectively. In both conditions, hepatic arterial flow into the involved hepatic parenchyma tends to increase, resulting in arterial enhancement during early phase dynamic MRI. Regarding the extent of the wedge-shaped enhancement, in cases of portal vein compromise, the involved area can be limited to the distal area from the compromised vascular branch [25]. In cases of hepatic venous obstruction, the extent of the arterial enhancing area is more variable, depending on the perfusion pressure and the total amount of blood flow [25, 27–29]. The involved area can be nonsegmental or rather geographic in cases of hepatic venous obstruction [27].
In addition to the simple mass effect of larger tumors, direct tumor cell infiltration or desmoplastic reaction in the adjacent portal vein or hepatic vein branches is always possible regardless of the tumor size, which could explain many exceptional cases of small lesions accompanying a profound wedge-shaped enhancement. Thus, even in wedge enhancement, the presence of perilesional enhancement implicates the possibility of tumor infiltration along the extralesional vascular pedicle [30], which influences therapeutic planning and prognosis. At this point, disappearance or decreased extent of perilesional wedge enhancement could be a positive sign of therapeutic effect after chemotherapy [11, 30].
This study has several limitations. There was no histologic proof for the perilesional enhancement area. As in previous studies [11–13], however, histologic investigation would not be the right tool to evaluate this kind of localized hemodynamic phenomena around the lesion. No one actually knows the precise size of hepatic tumors even if pathologic specimens are provided because there is an inherent difference between in vivo tissue and extracted ones and because of the technical difficulty of a direct comparison of right sectional planes for this kind of study. No other choice was available but to estimate the tumor size on the precontrast image to compare the variable thickness of rim enhancement on the postcontrast images with the same imaging parameters of precontrast images at the same or closest slice level in this study. Limited spatial resolution can be another obstacle to measuring the thickness of rim enhancement. Because the thickness of the enhancing rim was not always even and not always adequate to precisely measure at the submillimeter level, the subtracting method that we used might be more helpful than direct measurement to estimate the thickness of the rim enhancement by summation effect of both sides, subsequently divided into one half to close to a mean thickness.
Despite these limitations, our method could produce several statistically significant resultant values related to the perilesional enhancement of hepatic metastases with the tumor size or vascularity during the arterial phase dynamic MRI. Knowledge of the imaging features and an understanding of the mechanism of perilesional enhancement around hepatic metastases may result in more sophisticated access to the tumoral and epitumoral environment for therapeutic planning and an expectation of a better prognosis. This knowledge should also help radiologists more accurately characterize tumors.
Acknowledgments
We thank Prof. Osamu Matsui and Dr. Kazuhiko Ueda in Kanazawa, Japan, for
early discussion that contributed to the setup of the study concept.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  G. Jia, C. O'Dell, J. T. Heverhagen, X. Yang, J. Liang, R. V. Jacko, S. Sammet, T. Pellas, P. Cole, and M. V. Knopp Colorectal Liver Metastases: Contrast Agent Diffusion Coefficient for Quantification of Contrast Enhancement Heterogeneity at MR Imaging Radiology, September 1, 2008; 248(3): 901 - 909. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Catalano, F. Sandomenico, A. Nunziata, M. M. Raso, P. Vallone, and A. Siani Transient Hepatic Echogenicity Difference on Contrast-Enhanced Ultrasonography: Sonographic Sign and Pitfall J. Ultrasound Med., March 1, 2007; 26(3): 337 - 345. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
