| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Radiology, Beth Israel Deaconess Medical Center and Harvard Medical School, 1 Deaconess Rd, West Campus 302B, Boston, MA 02215 (J.B.K., R.A.K., S.N.G.); and Laboratory of Surgical Biology, Boston University School of Medicine, Mass (P.T.). Received February 4, 2003; revision requested April 14; final revision received September 15; accepted October 8. Supported by National Cancer Institute grant R21-CA89634–02. J.B.K. supported by the RSNA R & E Foundation through the RSNA Scholar Award. Address correspondence to J.B.K. (e-mail: jkruskal@bidmc.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Intravital imaging of exteriorized livers was performed in 72 live mice. Three groups of mice were studied: a sham-operated control group (n = 24), a group with nonmetastasizing subcutaneous gliomas (n = 24), and a group with developing hepatic CX-1 colon cancer metastases (n = 24). Microvascular flow parameters, leukocyte-endothelial interactions, and wall shear stress were directly measured in hepatic sinusoids and postsinusoidal venules at 2-day intervals prior to and during the development of metastases. The Kruskal-Wallis test was used initially to test for overall equality of medians in each data group. Single posttest comparisons of independent samples were performed with the Mann-Whitney test, with an overall statistical significance of .05.
RESULTS: Prior to the development of visible colorectal cancer metastases, significant (P < .05) reductions occurred in sinusoidal and postsinusoidal flow and wall shear rates, coupled with increased leukocyte rolling and adherence. With tumor growth, flow was further compromised in 92% of tumors larger than 0.5 mm in diameter by extrinsic compression of sinusoids and portal venules and narrowing caused by adherent leukocytes.
CONCLUSION: Significant intrahepatic flow alterations occur in mouse livers prior to growth of visible metastases and provide a rational explanation for elevation in the Doppler perfusion index that occurs prior to tumor formation.
© RSNA, 2004
Index terms: Animals • Colon neoplasms, 75.321 • Liver neoplasms, metastases, 761.332 • Ultrasound (US), Doppler studies, 76.12984
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Currently, the cause of perfusion changes is believed to be a humoral mediator-induced portal venous flow reduction (6) rather than an intrinsic hepatic hemodynamic event. Several important questions have been raised because of hepatic perfusion data, including the biologic relevance and precise cause of these flow alterations (7). To our knowledge, the intrahepatic flow and cellular alterations that occur during establishment of hepatic metastases have not been studied previously. The availability of an animal model withhepatic colorectal cancer metastases and the development and improvement of in vivo video microscopic techniques now make it possible to document and study these flow alterations in live animals prior to and during the formation of hepatic metastases. The purpose of our study was to evaluate whether intrahepatic flow alterations occur during formation of hepatic colorectal cancer metastases and to identify possible causes of these alterations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Animal Models
Experiments were performed according to a protocol approved by the Institutional Animal Care and Use Committee at Beth Israel Deaconess Medical Center, in accordance with the guidelines issued by the National Institutes of Health for the care of laboratory animals. All mice were housed in a pathogen-free temperature-controlled environment and were allowed access to food and water ad libitum. Inbred 6-week-old male athymic nude mice (Harlan-Sprague-Dawley, Indianapolis, Ind) that weighed 20–25 g were used in this study. Surgery, tumor cell injection, and optical microscopy were performed by one author (J.B.K.) following induction of anesthesia with an intraperitoneal mixture of ketamine (Abbott Laboratories, North Chicago, Ill) and xylazine (Abbott Laboratories) at a strength of 100 mg/mL, given as a dose of 0.1 mL per 20 g.
Three groups of mice were studied. Group I consisted of healthy sham-operated control mice (n = 24). Group II consisted of mice that received a subcutaneous inoculation of 2 x 106 nonmetastasizing subcutaneous glioma cells (n = 24). Group III consisted of mice that received an intrasplenic inoculation of 2 x 106 human CX-1 colon cancer cells (n = 24).
Mice in group III received a single intrasplenic inoculation of human CX-1 colorectal cancer cells, according to well-established protocols (8–10). Cell lines were checked frequently for the presence of mycoplasma with tissue staining (33258; Hoechst, Frankfurt, Germany), and the results were negative. Human CX-1 colon carcinoma cells (2 x 106 cells; 0.5 mL of phosphate-buffered saline, pH 7.4) with a moderately to well-differentiated human colorectal cancer line were derived from the HT-29 cell line and provided by L. B. Chen, PhD (Dana-Faber Cancer Institute, Boston, Mass). These cells were maintained in a medium (RPMI 1640; Gibco Laboratories Life Technologies, Grand Island, NY) with 10% fetal bovine serum (Sigma Chemical, St Louis, Mo), 1% L-glutamine (Gibco), 100 units/mL penicillin G (Gibco), and 100 mg/mL streptomycin (Gibco). Under direct vision, cells were injected into the spleen via a minilaparotomy incision by using a 27-gauge needle. These cells spread to the liver, and invasive and vascular metastases were established by day 7.
Mice in group II received a subcutaneous inoculation of nonmetastasizing H247 glioma cells (2 x 106 cells; 0.5 mL phosphate-buffered saline, pH 7.4) that was provided by L. B. Chen, PhD (Dana Farber Cancer Center). This model was selected because these cells do not metastasize to the liver and do not express a carcinoembryonic antigen, which is an integrin implicated in activation of hepatic Kupffer cells.
Hepatic in Vivo Optical Microscopy
Livers were studied at 2-day intervals for up to 8 days (five time points). Dynamic in vivo microscopy of exteriorized livers was performed in all mice at each of these time points.
After induction of anesthesia, a 2.0-cm vertical midline incision was made that extended along the left subcostal margin, and the liver was mobilized by incising the falciform ligament off the diaphragm. Mice were then placed in the left lateral position, and the exteriorized peripheral lobules of the livers were placed on an angled quartz crystal on a modified microscopy stage and transilluminated with a cooled monochromatic (500–800-nm) halogen light source. The segment of liver to be studied was covered with moist plastic film (Saran Wrap; Dow Chemical, Midland, Mich) to limit movement caused by respiration, cardiac motion, or peristalsis. At no point were livers handled manually; all hepatic manipulations were performed with cotton-tipped applicator sticks. Livers were kept warm and moist by means of continuous flushing with warm (37°C) lactated Ringer solution. The color of the mesenteric vessels was monitored regularly to confirm adequate respiratory and circulatory status, and no cyanosis was documented.
A standard compound trinocular microscope was modified for in vivo microscopy and equipped for transillumination and epiillumination. Images were obtained at a magnification of x100–1,500 with a modified microscope system (Optiphot; Nikon, Tokyo, Japan), and both color (SSC-DC50A Exwave CCD; Sony Medical Systems, Tokyo, Japan) and black and white (CCD-72; DAGE/MTI, Michigan City, Ind) images were recorded. Images were transferred to a computer (Optiplex; Dell Computers, Round Rock, Tex) by using a real-time video frame grabbing and processing graphics accelerator card (Radeon All-In-Wonder 8500 DV; ATI Technologies, Santa Clara, Calif) for use with image analysis software (NIH Image 1.62; National Institutes of Health, Bethesda, Md; available at: ftp://rsbweb.nih.gov/pub/nih-image/). When necessary, images were also stored on digital disks by using a recorder (miniDV; Sony) for subsequent data and image analysis.
Microvascular Flow Parameters
Microvascular blood flow rates (Qb) were determined optically in hepatic sinusoids and postsinusoidal venules from the product of mean red blood cell velocity (Vmean), which was derived from center-line flow velocity (Vc/1.6) (Baker-Weiland factor) and microvascular diameter (D) (11). These data were obtained from digitized 30-second video loops (1/29 second frame-rate intervals) (11) and were determined by one author (J.B.K.). Mean velocities were estimated by determining individual velocities of 25 consecutive noninteracting erythrocytes (mean, 26 erythrocytes ± 4; range, 14–31 erythrocytes) by measuring the distance traveled between two or more successive video frames. Each of these 25 measurements was obtained in 25 different sinusoids and postsinusoidal venules, respectively. For each of these velocity determinations, the corresponding vessel diameter was also recorded. From Vc, Qb was computed as follows: Qb = Vc/1.6 · 
· D2/4 · 10–6 µl/sec. Data were collected from 10 different peripheral lobules in each liver and from visible tumors in mice in group III only. These data included microvascular perfusion indexes, wall shear rates, and endothelial cell studies, as will be described later. Intratumoral hemodynamic parameters were measured in vessels containing flow within the CX-1 colon cancer metastases and were compared with flow in adjacent lobules from the same liver where no tumors were visible. These observations were made 8 days after intrasplenic inoculation of tumor cells to document whether these contributed to intrahepatic flow alterations.
Leukocyte-Endothelial Interactions
Leukocyte-endothelial interactions were studied in both the sinusoids and the postsinusoidal venules. To allow visualization of leukocytes, mice were injected with a 20-µL bolus of 0.1% rhodamine 6G (Sigma Chemical) in 0.9% saline (12), which is specifically taken up by circulating leukocytes. Parameters studied included rolling (or temporary adherent) and stagnant sinusoidal leukocyte counts, which are defined as leukocytes located within sinusoids containing moving flow that were adherent for less than 20 seconds or did not move for 20 seconds, respectively (13). In postsinusoidal venules, adherent leukocytes and leukocyte velocity were determined by a single author (J.B.K.). The adherent leukocyte counts are given per endothelial surface area and, assuming a cylindrical geometry, were calculated with the following equation: x D x observed length of vessel (13). The newtonian wall shear rate, 
w, was estimated in postsinusoidal venules as w = Vc/1.6 x 1/D x 8 (sec–1) (14).
Statistical Analyses
Data are given as mean ± standard error of the mean (SEM). Mean values of microcirculatory data represent the mean of all animals in each group at each time point calculated from the mean value of each animal. Data were initially assessed for normality with use of normal probability plots. On the basis of these results, it was decided to use nonparametric procedures to compare data from the three groups. Given the independent and unpaired nature of all data, the Kruskal-Wallis test was initially used to test for overall equality of medians in each data group. When statistically significant differences occurred, single posttest comparisons of independent samples were performed by using the Mann-Whitney test with an overall significance of .05. A difference was considered to be statistically significant if the P value was less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
|
|
Wall shear rate decreased with time in mice with developing metastases (Fig 9), indicating that leukocyte adherence to venous endothelium may be partly because of changes in local wall shear conditions. By day 4, the reduced shear rate in group III was significantly lower than the rate in groups I and II (P < .05).
Tumor Microhemodynamics
When compared with parameters in adjacent lobules, mean red blood cell velocities and flow rates, the proportion of stagnant and rolling leukocytes, and the wall shear rates were all reduced in tumor vessels (Table 2). The mean tumor vessel diameter on day 8, which was determined from 75 vessels within 27 CX-1 tumors, was 11.1 µm. The slowest flow occurred in nondistensible small (mean, 8.3 µm ± 1.1) tumor-penetrating vessels, which arose from peritumoral sinusoidal spaces. Fourteen (92%) of 15 tumors that were 0.5 mm in diameter produced extrinsic narrowing of portal venules or centrilobular sinusoids. This narrowing of vessels was further enhanced by increased leukocyte adherence in outflow venules, whereas the rate of leukocyte-endothelial interactions was lower in tumor vessels than that in adjacent perfused sinusoids.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Cause of Flow Alterations
This study focused on perfusion alterations that occur prior to growth of intrahepatic metastases. The precise cause of these changes remains conjectural. Humoral mediators have been implicated in altered hepatic perfusion (7), but to our knowledge, the nature or cell of origin of these mediators is not currently known. By injecting plasma samples from patients with and patients without overt colorectal metastases into rats, Warren et al (6) demonstrated splanchnic hemodynamic changes and postulated that a humoral mediator was present. In a rat model with hepatic metastases, Hemingway et al (15) showed a reduction in portal flow and an increase in portal resistance during early growth of hepatic micrometastases. The presence of a circulating vasoactive agent has been suggested, as a significantly (P < .05) increased vascular resistance developed in rats in which small-bowel segments were cross-perfused with arterial blood from tumor-bearing rats (16). Our demonstration of different flow parameters within tumors and adjacent sinusoids provides evidence for additional intrahepatic causes of these changes. Results of studies have shown that tumor cell activation of Kupffer cells results in cytokine release and endothelial cell activation (17). Circulating leukocytes bind to activated endothelial cells; this activation and binding may explain our observation of increased leukocyte rolling and stasis in livers with developing metastases. Interestingly, inhibition of receptors expressed on activated endothelial surfaces abrogates formation of metastases (18).
Clinical Relevance of an Altered Doppler Perfusion Index
For patients with hepatic metastases from colorectal cancer, surgery is currently the only option that offers the potential for cure (19). Earlier detection of hepatic metastases may improve survival, but there is currently no established method for predicting which patients are likely to experience a recurrence of colorectal cancer. Studies have shown that overt and occult hepatic metastases are associated with subtle changes in hepatic perfusion that can be detected by using Doppler ultrasonography (US) (4,20). The hepatic perfusion index predicts a poor outcome in patients with colorectal cancer and may be useful in patient selection for adjuvant chemotherapy (5). Leen et al (3) and Warren et al (5) have suggested that administration of adjuvant chemotherapy should be based on the Doppler perfusion index rather than on Dukes stage, which is the reference standard against which other prognostic factors are currently compared. Unlike Dukes stage, the Doppler perfusion index can be used to predict which patients are likely to experience a recurrence of colorectal cancer (5,21).
Specificity of Flow Alterations
The hepatic flow alterations observed in our study may be specific for colon cancer. Enhanced leukocyte adherence to endothelial surfaces was significantly less in mice with nonmetastasizing glioma cells (group II) than that in mice with developing hepatic metastases (group III). Cell membrane molecules such as the carcinoembryonic and the sialyl Lewis X antigen are present in certain colon cancer cells, including CX-1 cells. These molecules bind specifically to a Kupffer cell receptor and to E-selectin expressed by activated endothelial cells, respectively. We speculate that these specific cell membrane interactions are responsible in part for the specificity of the flow changes we have observed. In a computed tomographic (CT) study of patients with breast cancer, no significant hepatic perfusion differences were identified between those patients who developed subsequent hepatic metastases and those who did not (22). Since the Doppler perfusion index is more sensitive than CT to flow changes (23), it is possible that subtle changes may have been underestimated with CT.
Alterations in Shear Rate
We have shown a significant reduction in wall shear rate both in livers that are developing metastases and in tumors themselves. The velocity with which ligands contact endothelial selectins is important to increase the probability of collision. An increased collision rate enhances temporary and permanent binding of cells to a vessel wall. Our data provide an explanation for increased cell adhesion to vessel walls in livers prior to establishment of metastases. To resist substantial wall shear stress exerted by blood flow, metastasizing colon carcinoma cells have to form adhesive contacts with endothelial cells and subendothelial extracellular matrix (24). Assuming laminar flow, the wall shear rate reflects the velocity of fluid relative to adjacent layers. More precisely, the shear rate is the velocity gradient perpendicular to the direction of fluid flow. Formation of transient receptor-ligand bonds between circulating cells and the endothelial lining occur with an increase in the shear rate and do not form above cutoff flow rates (25). Chen and Springer (25) have shown that maximum tethering of neutrophils to selectin occurs at a shear rate of 100 sec–1. The reduced shear rate that we have documented, which shows a trend toward similar values, suggests an increasing likelihood of interactions between leukocytes (and presumably tumor cells) and endothelial receptors during formation of metastases. This is further supported by our data, which indicate increased leukocyte rolling and adherence in livers preceding tumor formation.
Leukocyte-Endothelial Interactions
Investigation of leukocyte-endothelial interactions with intravital microscopy allows evaluation of the expression and function of leukocyte adhesion molecules in the endothelium (26). In the present study, we have observed reduced leukocyte adherence in tumor vessels, enhanced leukocyte adherence in postsinusoidal venules, and leukostasis in hepatic sinusoids. The diminished interaction between leukocytes and tumor microvascular endothelium is well recognized (27) but to our knowledge has not been described with the colon cancer cell line used in our study. It is possible that adhesion receptors are either down-regulated or not adequately expressed by immature tumor endothelium. Postsinusoidal leukocyte adherence increased with duration of tumor growth. It is possible that with tumor growth, an increased number of tumor cells produced more humoral substances and resulted in this effect.
The enhanced leukocyte-endothelial interactions we have documented are not specific for tumor formation. Intrahepatic expression of selectins following hepatic ischemia and/or reperfusion is accompanied by increased leukocyte-endothelial cell adhesion in terminal hepatic venules (28). It is possible that different receptors are involved in cell adherence in different parts of the liver. For instance, intercellular adhesion molecule 1 is involved in the process that mediates leukocyte adherence in postsinusoidal venules, whereas other mechanisms are involved in the hepatic sinusoids (29). Our observation that the dominant interactions occur in postsinusoidal venules should help focus future research efforts on characterization of adhesion receptors at this site.
The cause of sinusoidal leukostasis is not currently known. We documented increased retention of leukocytes in patent sinusoids, with no increase in leukocyte rolling or adherence. This suggests an alternate mechanism that is distinct from endothelial up-regulation. Activated Kupffer cells play an important role in mediating impaired sinusoidal perfusion through the release of vasoactive cytokines. Since circulating CX-1 tumor cells activate Kupffer cells, we postulate that these activated Kupffer cells are releasing a vasoactive compound that may be responsible for altering sinusoidal perfusion.
We have recently identified an increased number of stellate cells in the livers of mice that are developing metastases. These cells are easily visible in vivo because of the presence of vitamin A, and their well-recognized contractile properties (30) may be responsible in part for some of the flow alterations we have identified. Gulubova (31) has shown increased numbers of myofibroblast-like transformed stellate cells in the sinusoids of patients with malignant tumors without hepatic metastases. When activated, these cells become contractile and moderate sinusoidal perfusion. Most interesting is that this activation occurs through Kupffer cell-released cytokines (32). It is likely that our circulating tumor cells activate Kupffer cells to release cytokines, which in turn activate stellate cells to modify perfusion. Given the potential relevance of these observations, we are actively exploring the precise role that these cells play in the formation of hepatic tumors.
Comparative Studies
Few published studies exist for making direct comparisons with our data. In a study that correlates Doppler US measurements with hepatic histologic findings in an animal model of hepatic metastases, Yarmenitis et al (33) demonstrated a distinct elevation of hepatic arterial flow during the early stage of metastasis formation with minimal reduction in portal flow. By using CT in a rat model of micrometastases, Cuenod et al (34) demonstrated a 34% decrease in portal blood flow with an increase in mean transit time through the liver. Platt et al (35) had similar findings with contrast material–enhanced CT. In several videomicroscopy studies (36,37), a variety of measurable indexes were used to evaluate tumor or hepatic blood flow. We have used similar indexes for obtaining measurements of flow, endothelial interactions, and shear rates in the present study, and our data compare favorably with the data of these studies, despite our use of different cell lines and a mouse model.
Study Limitations
This study has several factors that limit direct comparisons with the human clinical situation. First, it is possible that the flow alterations and endothelial interactions that occur in the mouse liver differ from those that occur in the human liver. We have, however, provided sufficient evidence to justify the study of similar events in human subjects. Second, we did not measure splanchnic or intrahepatic arterial or portal flow. Dynamic hepatic microscopy does not permit one to visualize morphology and function of hepatic arterioles and portal venules (13). Use of confocal imaging systems or whole-body optical devices may soon permit these observations to be made. Third, these flow changes may be specific for this cancer cell type. We are currently extending previous studies by using different colon cancer cell lines that possess different metastastic properties (38).
In summary, by using dynamic observations of intrahepatic microvascular perfusion prior to and during establishment of hepatic colorectal cancer metastases, we have confirmed the presence of and documented causes for reduced portal and hepatic venous perfusion. This experiment validates clinical studies of altered hepatic perfusion indexes in humans and provides a rational justification for further studies of these microvascular phenomena in patients.
Practical applications: This study has several important potential clinical applications. We have provided direct in vivo confirmation that specific intrahepatic hemodynamic changes do indeed occur prior to growth of hepatic metastases. These changes, which arise in part from alterations in leukocyte-endothelial adherence, suggest that up-regulation of endothelial receptors is occurring. It is possible that these receptors can be used for diagnostic, prognostic, or even therapeutic purposes. Indeed, peptides that inhibit E-selectin have recently been shown to inhibit formation of hepatic metastases (18), and the integrin inhibitor Eristostatin has also been studied (39). The precise cause of the reduced flow and endothelial up-regulation remains uncertain; however, it is known that CX-1 colorectal cancer–induced activation of Kupffer cells elicits expression and release of several cytokines that induce expression of endothelial E-selectin (40). In addition, activated Kupffer cells and hepatic stellate cells will express nitric oxide (30) and endothelin-1 (41), respectively, both of which modify hepatic perfusion. By using intravital imaging techniques, we are currently exploring the role that activated Kupffer and hepatic stellate cells play in modifying hepatic perfusion during formation of metastases.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, J.B.K.; study concepts, all authors; study design, J.B.K., P.T., S.N.G.; literature research, J.B.K., R.A.K.; experimental studies, J.B.K., P.T., S.N.G.; data acquisition, J.B.K., P.T.; data analysis/interpretation, all authors; statistical analysis, J.B.K., S.N.G.; manuscript preparation, J.B.K., S.N.G.; manuscript definition of intellectual content, editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. A. Miles, B. Ganeshan, M. R. Griffiths, R. C. D. Young, and C. R. Chatwin Colorectal Cancer: Texture Analysis of Portal Phase Hepatic CT Images as a Potential Marker of Survival Radiology, February 1, 2009; 250(2): 444 - 452. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J H ZHOU, A H LI, L H CAO, H H JIANG, L Z LIU, X Q PEI, and F HAN Haemodynamic parameters of the hepatic artery and vein can detect liver metastases: assessment using contrast-enhanced ultrasound Br. J. Radiol., February 1, 2008; 81(962): 113 - 119. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. B. Kruskal, A. Azouz, H. Korideck, M. El-Hallak, S. C. Robson, P. Thomas, and S. N. Goldberg Hepatic Colorectal Cancer Metastases: Imaging Initial Steps of Formation in Mice Radiology, June 1, 2007; 243(3): 703 - 711. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. B. Kruskal Can Optical Imaging Assist in Characterization of the Onset of Angiogenesis in Developing Hepatic Metastases in Mice Livers? Radiology, May 1, 2007; 243(2): 307 - 308. [Full Text] [PDF]
|
|
|
|
|
|
Y. Liu and O. Matsui Changes of Intratumoral Microvessels and Blood Perfusion during Establishment of Hepatic Metastases in Mice Radiology, May 1, 2007; 243(2): 386 - 395. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. O. Oktar, C. Yucel, T. Demirogullari, A. Uner, M. Benekli, G. Erbas, and H. Ozdemir Doppler sonographic evaluation of hemodynamic changes in colorectal liver metastases relative to liver size. J. Ultrasound Med., May 1, 2006; 25(5): 575 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C Harvey and M Blomley Imaging investigation of liver haemodynamics in patients at risk for hepatic metastatic disease Br. J. Radiol., February 1, 2005; 78(926): 103 - 104. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
