| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Experimental Studies |
1 From the Center for Molecular Imaging Research, Massachusetts General Hospital and Harvard Medical School, Bldg 149, 13th St, 5406, Charlestown, MA 02129. Received January 9, 2002; revision requested February 4; revision received April 2; accepted May 23. Address correspondence to R.W. (e-mail: weissleder@helix.mgh.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Tumors with differing degrees of angiogenesis—9L rodent gliosarcoma, DU4475 human mammary adenocarcinoma, HT1080 human fibrosarcoma, and EOMA hemangioendothelioma—were implanted in nude mice. Tumoral VVFs were measured at submillimeter voxel resolutions by using 1.5-T MR imaging. A technetium-labeled intravascular radiotracer was injected into a subset of the animals to validate the MR imaging measurements. Microvessel density and vascular endothelial growth factor (VEGF) also were measured. Statistical analysis was performed with analysis of variance.
RESULTS: High-resolution multisection MR maps of tumor blood volume were obtained in all tumor models. Mean tumoral VVF differed significantly among the different tumors: 2.1% ± 0.3 (standard error of mean) for 9L gliosarcoma, 3.1% ± 0.4 for DU4475 mammary adenocarcinoma, 5.5% ± 0.8 for HT1080 fibrosarcoma, and 6.6% ± 0.9 for EOMA hemangioendothelioma (P < .01). There was a strong correlation between the MR imaging and radiotracer measurements. There was considerable intra- and intertumoral heterogeneity among the VVFs. MR imaging measurements were in accordance with conventional measurements of angiogenesis, such as microvessel density count and VEGF.
CONCLUSION: Measurements of tumoral VVF at high-resolution MR imaging with long-circulating iron oxide are feasible and correlate with angiogenic burden in experimental tumor models.
© RSNA, 2002
Index terms: Angiogenesis • Contrast media, experimental studies • Experimental study • Neoplasms, experimental studies • Neoplasms, radionuclide studies
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The continued development of newer and more efficient angiogenesis inhibitors (5) has prompted the need to develop and validate imaging techniques to measure the physiologic and molecular surrogate markers of angiogenesis. In particular, imaging measurement techniques that are sensitive to tumoral microvessels (ie, vessels < 20 µm in diameter) and yield high spatial resolution for depiction of "hotbeds" of angiogenesis (ie, areas of extensive new vessel growth) are desirable. Because the tumors and metastases in patients involved in clinical trials are often multifocal and large (ie, end stage), such imaging techniques should also be capable of complete tumor coverage rather than limited single-section measurements. Rapid-sequence, single-section imaging measurements performed with low-molecular-weight contrast agents have been used to extract first-pass kinetics and derive mathematical parameters (6–10). Reliable multisection, steady-state, high-resolution imaging, however, requires the use of a blood-pool contrast agent—that is, one with which a constant concentration of the agent stays within the intravascular space during the entire time of imaging. Although a number of such agents have been used in experimental studies (11,12), few are likely to be developed for clinical use.
One of the few classes of clinically viable, truly intravascular imaging agents that are being developed for clinical use are the long-circulating dextrinated iron oxide (LCDIO) preparations for use with magnetic resonance (MR) imaging (13,14). Before these agents are adapted for use in imaging tumor vascularity, however, their effectiveness must be experimentally validated. Thus, a model LCDIO preparation that has a blood half-life of several hours in vivo and is similar to preparations used in clinical trials (15) needs to be tested. Therefore, the purpose of our study was to experimentally validate the effectiveness of MR imaging performed with LCDIOs in the measurement of vascular volume fractions (VVFs) as indicators of tumor angiogenesis in different experimental tumor models.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The 9L and EOMA cells were cultured in modified Eagle’s medium (Cellgro; Mediatech, Washington, DC) supplemented with 2 mmol/L of L-glutamine and Earle salt solution with 1.5 g/L of sodium bicarbonate and 10% heat-inactivated fetal bovine serum (Cellgro FBS; Mediatech) at 37°C in a humidified 5% CO2 atmosphere. The HT1080 cells were cultured in the same medium supplemented with 0.1 mmol/L of nonessential amino acids and 1.0 mmol/L of sodium pyruvate. The EOMA cells were grown in modified Eagle’s medium (Cellgro) supplemented with 2 mmol/L of L-glutamine and Earle salt solution with 1.5 g/L of sodium bicarbonate and 20% heat-inactivated fetal bovine serum (Cellgro FBS).
To grow and propagate the tumors in mice, 2 x 106 cells of each tumor cell line were implanted bilaterally into the gluteal region in 16 athymic female nude mice that weighed 25–30 g. A total of 32 tumors—eight tumors per tumor model—were implanted. The animals were examined with MR imaging when the tumors had grown to a diameter of 5–7 mm. The mean time between tumor implantation and MR imaging was 16 days for 9L, 13 days for HT1080, 21 days for DU4475, and 10 days for EOMA tumors. At the end of the MR imaging experiments, the animals were sacrificed by means of ketamine and xylazine overdose. All animal studies were approved by the institutional animal care committee at Massachusetts General Hospital.
MR Image Acquisition and Processing
All MR imaging studies were carried out by using a 1.5-T clinical MR imaging system (Signa; GE Medical Systems, Milwaukee, Wis) and clinically available pulse sequences. All tumor-bearing mice were anesthetized by means of intraperitoneal injection of ketamine (80 mg/kg) and xylazine (12 mg/kg). Custom-made 28-g catheters were inserted into the lateral tail veins and attached to a micro–heparin-saline flush unit. A 3-inch surface receiver coil was used for all studies. Following a fast spoiled gradient-echo localizer sequence (34.0/2.2 [repetition time msec/echo time msec], 30° flip angle), multiple transverse images of the tumors in each animal were obtained. All MR imaging examinations consisted of a conventional gradient-echo sequence with parameters that included 3,000/20, a 90° flip angle, a 128 x 128 matrix, and two signals acquired. The field of view was set at 12 x 6 cm, and the section thickness was 2 mm. All animals were imaged before and after intravenous injection of 5 mg of the LCDIO agent, monocrystalline iron oxide 46, per kilogram of body weight. In addition, at the end of each LCDIO-enhanced MR imaging examination, gadopentetate dimeglumine–enhanced (0.1 µmol/kg) (Magnevist, Berlex Laboratories, Wayne, NJ) spin-echo images (300/6) were obtained to outline the tumor.
Steady-state tumoral blood volume maps were calculated from the pre- and postcontrast LCDIO-enhanced MR images, as described in detail elsewhere (6,17). A fundamental assumption is that the change in the transverse relaxation rate (
R2*) relative to the preinjection baseline relaxation rate is proportional to the perfused local blood volume per unit tumor volume (V) multiplied by a function (f) of the plasma concentration of the agent (P): R2* = k · f(P) · V.
Assuming a steady state of LCDIO distribution, the equation is revised to express a simple linear relationship between the change in the transverse relaxation rate and the perfused blood volume fraction: R2*(t) = k · V(t) or V(t) = [R2*(t)]/k, where R2*(t) is the change in the transverse relaxation rate of the tumor, V(t) is the tumor volume, and the constant k includes the agent blood pool concentration and is therefore dose dependent.
The enhancement of transverse relaxation can thus be expressed as follows: R2* = [(1/T2*post) - (1/T2*pre)] 
(-1/TE) · [ln(Spost/Spre)], where S is the signal intensity; TE, the echo time; and T2*, the transverse relaxation time. On the basis of this formula, maps depicting the change in transverse relaxation rate were calculated from all the MR images by using commercially available software (MatLab 7.0; The Mathworks, Waltham, Mass). In addition, two of the authors (C.B., M.M.) selected regions of interest that included the entire tumor and the adjacent muscle tissue on a given section (20–100 pixels) to calculate tumoral VVFs. Absolute tumoral VVFs were obtained by scaling measurements to muscle with a known VVF of 1.89% (18).
Tracer Blood Volume Determination
To independently verify the MR imaging measurements, we also obtained global tumoral blood volume measurements in the mice by performing nuclear imaging of 10 9L, six DU4475, eight HT1080, and eight EOMA tumors. For these studies, the animals were intravenously injected with a model long-circulating intravascular probe, methoxy poly(ethylene glycol)-poly-L-lysine (PL)-diethylenetriaminepentaacetic acid (molecular weight, 500 kD), that had been labeled with technetium 99m (99mTc) (19,20). The mice were positioned so that one of the authors (M.M.) could draw select regions of interest over the tumors without measuring the blood volume of other tissues. Within 15 minutes after this intravenous administration, the animals were imaged with a modified whole-body, single-head gamma camera (M-Cam; Siemens Medical Systems, Hoffman Estates, Ill) for 1 x 106 counts. Region of interest analysis and subsequent biodistribution experiments yielded values of tumor count per gram of tissue, and, subsequently, VVF values, by means of scaling to muscle radiotracer distribution.
Intravital Microscopy
In a separate control experiment, we sought to determine whether LCDIO is truly intravascular within tumor beds during the time of normal MR image acquisition—that is, within 30 minutes after intravenous administration. For these studies, we used intravital microscopy and fluorescently (indocyanine 5.5; Amersham Biosciences, Piscataway, NJ) labeled monocrystalline iron oxide 46. For these studies, 106 9L tumor cells enhanced with green fluorescent protein were injected into the ear pinna of three nu/nu mice bred at Massachusetts General Hospital. Ten days later, when the tumors had reached a size of approximately 40 mm3, the mice were anesthetized. Subsequently, indocyanine 5.5–labeled monocrystalline iron oxide 46 (5 mg of iron per kilogram of body weight) was injected into the lateral tail vein. The ear pinna was fixed on a microscope glass slide and observed at epifluorescent microscopy with use of different filters and magnifications (x2–x60). Microscopic images were obtained in the green fluorescent protein (vessels appear dark on a green fluorescent protein fluorescent tumor background) and indocyanine 5.5 (vessels appear bright on a dark background) channels for up to 30 minutes after intravenous injection of the fluorescently labeled LCDIO preparation.
Histologic Microvessel Density Measurements
Microvessel density counting at histologic analysis, a much used reference standard in angiogenesis studies, was performed in all animals (21). For these measurements, the tumors were excised, snap frozen, cut into 8-µm sections, and stained for CD31. To identify vessels unequivocally, the sections were incubated with a primary anti-CD31 antibody (Santa Cruz Biotechnology, Santa Cruz, Calif) and revealed with an alkaline phosphatase–labeled secondary antibody (Sigma, St Louis, Mo). After inactivation of the endogenous alkaline phosphatase by means of heating (65°C for 30 minutes), specific alkaline phosphatase activity was visualized by using NBT/BCIP substrate (Boehringer-Mannheim, Indianapolis, Ind). The sections were counterstained with nuclear fast red. Control sections were processed identically, with the exception of omitting the incubation with the primary antibody. Microvessel density counting was performed independently by two investigators (C.B., M.M.).
Western Blotting for Vascular Endothelial Growth Factor Measurements
Vascular endothelial growth factor (VEGF) is one of the major angiogenesis-promoting factors, and measurements of this factor have served as surrogate markers of vascularity. For these reasons, one of the authors (M.M.) also determined the VEGF levels of the different tumor cell lines. Tumor cell lysates were centrifuged for 10 minutes at 16,000 rpm. The protein content of the samples was determined by using a bicinchoninic acid assay (Pierce, Rockford, Ill). Equal amounts of protein were applied to a 12% sodium dodecyl sulfate–polyacrylamide electrophoresis gel. The proteins were separated for 50 minutes at 120 V and then blotted onto a polyvinylidene fluoride membrane (Hybond C; Amersham, Piscataway, NJ). After blocking, the membranes were incubated with a primary anti-VEGF antibody (Santa Cruz Biotechnology) and then incubated with a biotinylated secondary antibody (Vector Laboratories, Burlingame, Calif). Binding of the antibodies was revealed by means of an alkaline phosphatase conjugate by using NBT/BCIP as a substrate. Levels of VEGF were determined by scaling the intensity of bands to known standards.
Statistical Analysis
Data are presented as means ± standard errors of the means. Statistically significant differences between groups were determined by performing one-way analysis of variance. CIs were corrected for multiple comparisons; a P value of 
.025 was considered to indicate a significant difference. To compare the VVFs measured at MR imaging with those measured by using radiotracer nuclear imaging, we chose the two-tailed Student t test. For this analysis, a P value of <.05 was considered to indicate a significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
All tumors had similar signal intensities at T1- and T2-weighted MR imaging and were clearly demarcated from the surrounding tissues. The bilateral tumors typically had a similar overall volume; however, section-to-section variations existed owing to slight differences in tumor implantation location (Fig 3). The global tumoral VVFs obtained in all animals are summarized in the Table. There was no statistically significant difference in global VVF values between the MR imaging and nuclear imaging (ie, radiotracer) measurements.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Nanoparticle-based iron oxides represent one of the few classes of contrast agents with a large hydrodynamic diameter (20–50 nm) and that have been developed for clinical imaging. LCDIOs are still small enough to circulate freely in the plasma, however, even in certain vascular beds where large cells and ultrasonographic bubbles are excluded (because they have large diameters, of several micrometers). Another important factor was the development of nanoparticles with extended blood half-lives (ie, no short clearance), which has enabled MR imaging measurements to be performed in steady state.
The initial development of LCDIOs began in the late 1980s with the observation that large dextran-coated iron oxide colloids contained a fraction of small particles, which were initially termed ultrasmall superparamagnetic iron oxides and had a long blood half-life—that is, of hours rather than minutes (14,23). Subsequent efforts to directly synthesize such small iron oxides at a higher yield were undertaken.
The structure and biologic properties of two LCDIOs synthesized with different methods, monocrystalline iron oxide (24,25) and ferumoxstran (Combidex; Advanced Magnetics, Cambridge, Mass), were initially developed independently (26,27). LCDIO (monocrystalline iron oxide or ferumoxstran) consists of 4–6-nm cores of iron oxide that are covered with a layer of T-10 dextran that is approximately 10–20 nm thick; this yields an overall volume (volume weighted) of aqueous solution of 20–50 nm, depending on the preparation. Both of these LCDIOs have a blood half-life of longer than 11 hours in mice, and they have similar biodistributions. On the basis of vascular enhancement at T1-weighted MR imaging, the blood half-life of Combidex in humans is longer than 24 hours. Phase 3 clinical testing of Combidex in the United States has been completed, and this agent is commercially available in Europe. As such, this agent could be used for angiogenesis studies in patients who are undergoing antiangiogenic therapy.
Our rationale for choosing LCDIO as the contrast agent for tumoral blood volume measurements is based primarily on four observations: (a) LCDIOs have long circulation times, which allow steady-state imaging (to image multiple tumors or perform additional MR measurements), (b) the sensitivity of T2-type agents for depiction of subtle changes in blood volume should greatly exceed that of the T1-based agents (10), (c) LCDIOs are long-circulating MR imaging contrast agents that do not readily extravasate into the tumor interstitium during imaging, and (d) LCDIOs are being developed for clinical use.
Although LCDIOs have been used extensively for mapping cerebral blood volumes in task activation functional MR imaging studies (10), the use of these agents for measuring tumoral blood volumes has been less explored, and this was our rationale in performing this study. A fundamental assumption in such measurements, which is supported by detailed numeric simulations, is that the change in transverse relaxation rate relative to the baseline transverse relaxation rate before contrast agent injection is proportional to the local blood volume fraction (see Materials and Methods). Similar techniques have been used in preliminary feasibility studies in patients (28).
The results of this study confirm the hypothesis that LCDIOs can be used to measure regional tumoral VVFs at steady-state MR imaging and to map the heterogeneity of VVFs among different tumors. As shown in the Table, the MR imaging measurements seem to be reasonably accurate compared with the other surrogate markers of angiogenesis, the nuclear (ie, radiotracer) measurements of global VVF in particular. MR imaging, however, was capable of yielding high-resolution maps of VVF distribution (Figs 3, 4), which may be particularly important in estimating the effectiveness of antiangiogenic agents, especially when they have heterogeneous effects.
Several differences between steady-state LCDIO-enhanced MR imaging and dynamic fast gadopentetate dimeglumine–enhanced MR imaging examinations of tumor vasculature are worth mentioning. First, LCDIOs represent a class of truly intravascular markers (with constant concentrations maintained during imaging) that allow steady-state imaging—that is, multiple tumors or metastases can be surveyed with a single injection of the agent. In contrast, gadopentetate dimeglumine–derived measurements are dependent on very fast MR imaging (ie, echo planar or fast gradient-echo imaging) over the same section for construction of time-activity curves that can then be used to mathematically derive parameters such as capillary permeability or surface area, blood flow, and, potentially, blood volume. One observation gleaned from intravital microscopic studies, however, is that there is marked variability in the intratumoral behavior of small molecules such as those of gadopentetate dimeglumine: Although some areas of the tumor exhibit first-pass leakage (with extraction fractions of up to 50%), other areas may exhibit much lower extractions. Predicting this behavior (as well as other parameters such as water exchange rates) mathematically may be difficult.
Finally, the various imaging methods have different sensitivities in the measurement of different physiologic parameters: Depending on the mechanism of action of a given angiogenesis inhibitor (eg, anti-VEGF action, antiendothelial action, and matrix-metalloproteinase inhibition), different approaches may be more sensitive than others. Results of intravital microscopic studies indicate that the functional tumoral VVF (ie, the perfused portion of blood within a tumor) represents a sensitive overall parameter of tumor angiogenesis.
Practical application: The described technique and its validation by means of MR imaging with a truly intravascular MR imaging radiotracer have considerable clinical applications in the measurement of tumor angiogenesis in vivo. Whether LCDIO-enhanced MR imaging is used to measure the effects of antiangiogenic agents in clinical trials or to determine the degree of vascularity in native tumors, the simplicity and validity of this examination are important aspects. Because similar agents are commercially available in Europe and will soon be available in the United States, they can be readily used in clinical trials. To make available a noninvasive method of imaging tumor angiogenesis at high spatial resolutions and repeatedly within the same patients has been the goal of much angiogenesis imaging research.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, R.W.; study concepts, C.B., R.W.; study design, A.B., C.B., R.W.; literature research, M.M., C.B., R.W.; experimental studies, V.N., C.B.; data acquisition, A.P., V.N., C.B.; data analysis/interpretation, V.N., M.M., C.B., R.W.; statistical analysis, C.B.; manuscript preparation, C.B., R.W.; manuscript definition of intellectual content, R.W.; manuscript editing, C.B., R.W.; manuscript revision/review and final version approval, R.W.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Persigehl, R. Bieker, L. Matuszewski, A. Wall, T. Kessler, H. Kooijman, N. Meier, W. Ebert, W. E. Berdel, W. Heindel, et al. Antiangiogenic Tumor Treatment: Early Noninvasive Monitoring with USPIO-enhanced MR Imaging in Mice Radiology, August 1, 2007; 244(2): 449 - 456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Pereira, D. E. Huddleston, A. M. Brickman, A. A. Sosunov, R. Hen, G. M. McKhann, R. Sloan, F. H. Gage, T. R. Brown, and S. A. Small An in vivo correlate of exercise-induced neurogenesis in the adult dentate gyrus PNAS, March 27, 2007; 104(13): 5638 - 5643. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Montet, J.-L. Figueiredo, H. Alencar, V. Ntziachristos, U. Mahmood, and R. Weissleder Tomographic Fluorescence Imaging of Tumor Vascular Volume in Mice Radiology, March 1, 2007; 242(3): 751 - 758. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Matuszewski, T. Persigehl, A. Wall, N. Meier, R. Bieker, H. Kooijman, B. Tombach, R. Mesters, W. E. Berdel, W. Heindel, et al. Assessment of Bone Marrow Angiogenesis in Patients with Acute Myeloid Leukemia by Using Contrast-enhanced MR Imaging with Clinically Approved Iron Oxides: Initial Experience Radiology, December 1, 2006; 242(1): 217 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Kelloff, K. A. Krohn, S. M. Larson, R. Weissleder, D. A. Mankoff, J. M. Hoffman, J. M. Link, K. Z. Guyton, W. C. Eckelman, H. I. Scher, et al. The Progress and Promise of Molecular Imaging Probes in Oncologic Drug Development Clin. Cancer Res., November 15, 2005; 11(22): 7967 - 7985. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. R. Kim, A. Yudina, J. Figueiredo, W. Reichardt, D. Hu-Lowe, A. Petrovsky, H. W. Kang, D. Torres, U. Mahmood, R. Weissleder, et al. Detection of Early Antiangiogenic Effects in Human Colon Adenocarcinoma Xenografts: In vivo Changes of Tumor Blood Volume in Response to Experimental VEGFR Tyrosine Kinase Inhibitor Cancer Res., October 15, 2005; 65(20): 9253 - 9260. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Tang, M. Kim, D. Carrasco, A. L. Kung, L. Chin, and R. Weissleder In vivo Assessment of RAS-Dependent Maintenance of Tumor Angiogenesis by Real-time Magnetic Resonance Imaging Cancer Res., September 15, 2005; 65(18): 8324 - 8330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kessler, R. Bieker, T. Padro, C. Schwoppe, T. Persigehl, C. Bremer, M. Kreuter, W. E. Berdel, and R. M. Mesters Inhibition of Tumor Growth by RGD Peptide-Directed Delivery of Truncated Tissue Factor to the Tumor Vasculature Clin. Cancer Res., September 1, 2005; 11(17): 6317 - 6324. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Montet, V. Ntziachristos, J. Grimm, and R. Weissleder Tomographic Fluorescence Mapping of Tumor Targets Cancer Res., July 15, 2005; 65(14): 6330 - 6336. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Matuszewski, T. Persigehl, A. Wall, W. Schwindt, B. Tombach, M. Fobker, C. Poremba, W. Ebert, W. Heindel, and C. Bremer Cell Tagging with Clinically Approved Iron Oxides: Feasibility and Effect of Lipofection, Particle Size, and Surface Coating on Labeling Efficiency Radiology, April 1, 2005; 235(1): 155 - 161. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Miller, H. H. Pien, D. Sahani, A. G. Sorensen, and J. H. Thrall Imaging Angiogenesis: Applications and Potential for Drug Development J Natl Cancer Inst, February 2, 2005; 97(3): 172 - 187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Krix, F. Kiessling, S. Vosseler, N. Farhan, M. M. Mueller, P. Bohlen, N. E. Fusenig, and S. Delorme Sensitive Noninvasive Monitoring of Tumor Perfusion during Antiangiogenic Therapy by Intermittent Bolus-Contrast Power Doppler Sonography Cancer Res., December 1, 2003; 63(23): 8264 - 8270. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
