| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
1 Departments of Radiological Diagnostics and Therapy and
2 Department of Carcinogenesis and Differentiation, German Cancer Research Center, Heidelberg, Germany, and
3 ImClone Systems, New York, New York
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this respect, radiological studies provide an ideal basis because they are entirely noninvasive and therefore almost universally applicable. However, measuring perfusion (blood flow/tissue unit) quantitatively is a considerable problem with all diagnostic imaging modalities, even including MRI.4 T2*-weighted dynamic susceptibility imaging is suitable to measure perfusion (9) but only where an arterial input function can be obtained and as long as contrast media remain in the intravascular space. The use of experimental magnetic resonance blood pool agents has shown encouraging results to improve blood volume measurement (10) . With T1-weighted dynamic MRI, suitable parameters of blood volume can be assessed in experimental tumors (11) ; however, these are only indirect measures based on the extravazation of contrast agent into the interstitial space.
In contrast, US techniques show encouraging possibilities to directly evaluate vessel perfusion during angiogenesis (12)
. Power Doppler sonography can detect reductions of larger tumor vessels during antivascular therapy (13, 14, 15, 16)
. However, conventional and even contrast-enhanced Doppler US is not capable of visualizing capillary blood flow (17)
, an essential precondition for quantifying tissue perfusion. Therefore, monitoring therapies that mainly affect the microvascular compartment may suffer limitations with these US methods. Potentially as a consequence of this lack of microvessel detection, in most previous studies, no correlation was observed between US parameters and the histologically determined vessel density (16
, 18, 19, 20, 21)
. There are, however, newly developed US methods that might be suitable to circumvent these limitations. High-frequency US is a novel technique, capable of detecting therapy-induced changes in small vessels with a diameter down to 
50 µm. However, its use is restricted to superficial tissue (<10 mm, malignant melanoma; Ref. 22
).
Additionally, intermittent sonography is a novel approach (23) that uses replenishment kinetics of a US contrast agent. Other than conventional US, it allows to indirectly assess capillary blood flow and is thus well suited to quantify functional parameters of tumor vascularization, including perfusion. The ability of this technique to quantify blood flow has been documented in previous studies where parameters derived by intermittent sonography were correlated with various measurements ex vivo [explanted veins (23) , kidney perfusion model (24) ], in vivo [myocardial blood flow measurement with radiolabeled microspheres (23 , 25 , 26) ], and in phantoms (23 , 27) . Several clinical studies on quantifying tissue perfusion of organs have been described in cardiology, neurology, or nephrology (28, 29, 30) . Limitations are related to general restrictions for US examinations (e.g., motion artifacts and energy loss in deeper tissue regions). Recently, we have further developed IBS and, for the first time, demonstrated that this method enables the quantification of tumor vasculature even in small animals (31) . The calculated mean blood velocity in tumors of 0.3 mm/s was within the expected magnitude.
In this study, we used this newly developed IBS in small experimental animals to successfully monitor quantitative alterations in the vascularization induced by systemic antiangiogenic therapy.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Five of the mice were treated with 800 µg of a monoclonal antibody to VEGF receptor 2 (DC101; ImClone Systems, New York, NY) every other day by i.p. injection, starting at day 21 postimplantation when tumors had reached a size of 0.3 ml (after the first US examination). This antiangiogenic therapy had successfully been tested previously in a comparable model (34)
.
Examination Method: Intermittent Sonography.
Intermittent sonography followed a single bolus injection of 100 µl of galactose-based US contrast agent (300 mg/ml Levovist; Schering, Berlin, Germany) injected within 5 s into a tail vein. The principles and theoretical background of the IBS have been described in detail elsewhere (23
, 31)
. After destruction of the intravasal contrast agent (microbubbles) by a high-energy US pulse, it is possible to derive replenishment kinetics of the microbubbles in an individually chosen ROI, as long as microbubbles are delivered from outside the ROI by the systemic circulation. As high US signals can be measured with high sensitivity (already one microbubble can be detected even from stationary microbubbles), capillary blood flow can be measured when microbubbles are given enough time to fill the small vessels (long pulsing interval between US pulses). According to the model of Wei et al. (23)
, the replenishment curve (US signal intensity over the time after destruction of microbubbles) describes an exponential increase, followed by a saturation behavior. The plateau A of this curve is proportional to the total blood volume of the examined region. The parameter ß indicates the increase of the curve and allows the approximation of mean blood velocity v in mm/s if the US beam width d is known (v = ß x d). In the used setting (see "Examination Protocol"), the beam width had been previously determined to be 0.78 mm (31)
.
For examining small animals with intermittent sonography, a single bolus injection is needed instead of the usual continuous infusion because reliable continuous infusions of the required small volumes and flows of contrast agent are hardly ever achievable. Consequently, the additional time-dependent decrease of the systemic microbubble concentration after the initial peak has to be considered. Nevertheless, it is possible to calculate replenishment parameters by calculating relative values of A dependent on a main pulsing interval (Fig. 1)
and including the maximum (max) of the US signal intensity-time curve after contrast agent injection (31)
. Thus, according to the model of intermittent sonography, parameters can be derived, which are proportional to blood volume (B max x A), blood flow [f A x max x ß, (ml x s-1)], and perfusion [blood flow/tissue unit, P A x max x ß x V-1, (ml x s-1 x mg-1)] of a tumor with the volume V.
|
.
|
. Three weeks after tumor induction (20 ± 1 days), IBS examinations of 10 nude mice were performed, and each animal was repeatedly examined every week, throughout 1 month. Five were treated with the VEGF receptor antibody, and 5 were kept as control animals.
|
l x w x d x 
/6, with l being the maximal length, w the width, and d the depth of tumor. Because the initial blood flow in the tumors’ groups (controls and treated animals) differed, values were normalized to the median value of initial blood flow f (n.a.u.) and perfusion P (n.a.u.), respectively. Animals were anesthetized by i.p. injection of ketamin (0.1 mg/g body weight, Ketanest; Parke-Davis, Berlin, Germany) and xylazine (1.5 ng/g body weight, Rompun; Bayer Vital, Leverkusen, Germany). The animal experiments were officially approved and complied with legal requirements and institutional guidelines.
Examination Protocol.
A Siemens-Acuson Sequoia 512 (Erlangen, Germany) US scanner was used, with a fixed linear array transducer 15L8w (power Doppler transmit frequency of 7 MHz). US device parameters were: maximum power (mechanical index, 1.9); maximum pulse repetition frequency (scale, 0.55) to reduce motion and noncontrast flow artifacts; high spatial resolution ('S2'); and standard for all other device parameters. The size (2 x 2 cm) and depth (>0.5 cm) of the color Doppler box, as well as the focus depth (tumor center), were kept constant; the main frame rate was 1.33 s-1. Intermittent imaging was started 1 min after the bolus injection of Levovist and lasted 380 s. This starting point and the following time interval had proven suitability in our previous study (31)
. A commercial software (Data Pro; Noesis, Courtaboeuf Cedex, France) was used for quantifying the number of color pixels (power Doppler mode), assuming these values were proportional to the number of microbubbles inside the ROI (35)
. For calculating the mathematical fit of the replenishment curves, we used nonlinear least squares fitting.
In addition, the data were calculated for two ROIs. The first was drawn around the entire tumor circumference, the second ROI was positioned in the central tumor region in a distance from all tumor borders of 2–3 mm. In tumors with a volume <0.2 ml (at the beginning and when tumors had regressed during therapy), such discrimination was not possible. In treated tumors, the US video signal in the central tumor parts was very low. Therefore, to avoid substantial calculating errors of the nonlinear last squares fitting in these tumors, the central blood flow was calculated by subtracting the peripheral from the total blood flow. The periphery was defined by subtracting the central ROI from the total ROI.
All 10 mice were repeatedly examined (each animal for three to four times). In 1 animal, the first i.v. injection failed. Four untreated mice had to be sacrificed after the third examination according to animal protection legislation because of large tumor size. One treated mouse died after the third examination for unexplained reasons. Thus, only results until the third of the weekly examinations were used for statistical evaluation.
Histology.
For histological diagnosis and assessment of vital tissue and its vascularization, eight tumors were dissected at given intervals: three untreated tumors 2 weeks after start of treatment and four treated tumors and one untreated tumor after 3 weeks of treatment (Fig. 2)
. Dissected tissues were covered with tissue tek (Sacura, Zoeterwoude, the Netherlands) and frozen in liquid nitrogen vapor. The tumor slice, which had been examined with IBS a few minutes before, had been tagged with tattoos (s.c. placed ink dots) to match scanning and cutting plane as well as possible 5-µm thick sections were cut with a Reichert-Jung Frigocut 2700 microtome. For morphology, sections were stained with H&E. Blood vessels and tumor cells were visualized by double immunofluorescence staining using a rat monoclonal antibody against PECAM (anti CD 31; PharMingen, San Diego, CA) and a guinea pig pankeratin antiserum (Progen, Heidelberg, Germany), respectively, following a standard protocol (34)
.
Tissue sections were viewed by phase contrast and fluorescence microscopy using an Olympus AX-70 microscope. Images were captured with an analysis SIS color view 12 digital camera (Soft Imaging System), and morphometric analyses were performed at x100 magnification using the analysis SIS (Soft Imaging System) software. To quantify vessel density, including microvascularity, the fraction of stained vessels/square millimeter was determined over the whole section, as well as in the central region.
Statistical Evaluation.
Changes of tumor parameters during therapy compared with those of untreated tumors were evaluated using Mann-Whitney-Wilcoxon tests (U test) with a significance level at P < 0.05. Pearson’s correlation coefficient was used to compare vessel density with IBS parameters and changes of tumor volume with perfusion or blood flow, respectively.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The initial median tumor volume in the control group was 0.34 ml (range, 0.004–0.42 ml) and increased to 0.63 ml (range, 0.01–0.98 ml) 2 weeks later. Conversely, the median tumor volume in the treatment group was reduced after 2 weeks under therapy from initially 0.30 ml (range, 0.01–0.47 ml) to 0.08 ml (range, 0.02–0.29; P < 0.05).
|
and 5
). Nondynamic US pictures at the time of maximal enhancement after contrast agent injection could visualize individual variations of tumor perfusion before start of therapy (Fig. 4, A and C)
. Particularly in small tumors, high vascularization signals could be observed (Fig. 5A)
.
|
|
. Although there was some initial variation in vascularization visible before treatment (Fig. 4, A and C)
, the contrast-enhanced power Doppler signals were clearly reduced in the treated animals, whereas they additionally increased in the untreated animals (Fig. 4D)
. Importantly, a reduction in tumor vascularization could already be observed in tumors under therapy before a reduction in tumor size became visible (Fig. 5)
.
Correlation with Vessel Density.
At histological examination, the calculated vessel density in the examined ROIs ranged from 0.27 to 2.60%. An untreated tumor with high vascularization signals in both the immunofluorescence staining and the corresponding US picture at the time of maximum enhancement after contrast agent injection is depicted in Fig. 6, A and C
. In contrast, in the treated tumor, histology as well as US data showed clearly lower signals (Fig. 6, B and D)
. Quantitatively, the vessel density on the microscopic slice (whole tumor) correlated well (k = 0.84, P < 0.01; Fig. 7
) with the blood volume density calculated from intermittent sonography (Table 1)
. When ROIs covering only the central part of the tumors were included in the analysis, this correlation was even more pronounced (k = 0.79, P < 0.001). Interestingly, treated tumors did not show significantly lower overall vessel densities than controls, whereas there was an excellent correlation between blood volume density and vessel density. Because of the large tumor size and thus development of necrotic areas at late stages, vessel density was reduced also in untreated tumors at the analyzed end points of the experiment (2 and 3 weeks after start of therapy, respectively). Dynamic parameters determined by intermittent sonography such as blood flow, blood velocity, or perfusion showed no correlation with the vessel density.
|
|
). Despite the randomization of the tumors according to pretherapeutic size to either the treatment group or the controls, the treatment group showed higher initial blood flow values (range, 1.6–6.4 a.u.) than the control group (range, 0.8–3.0 a.u.). However, this bias should not substantially interfere with the results of treatment effects because it were the treated tumors, which initially showed the higher vascularization.
|
and 5
): tumor blood flow f in tumors treated with VEGF receptor antibody was different from that in the controls (Fig. 8)
. In the control group, blood flow increased 1 week after start of observation (median increase, 409%) and then remained high. In the treated tumors, it decreased (median reduction, 64%) and remained low throughout the whole treatment and observation period (P < 0.05). Tumors in the treatment group with an initially higher blood flow showed a larger relative decrease in tumor size (
V/V) during the following 2 weeks under therapy than did tumors with a lower blood flow (k = -0.90; P < 0.05).
|
. However, there were high interindividual differences in mean blood velocity.
Furthermore, the blood volume B increased in untreated tumors during the observation period of 2 weeks but remained low and slightly decreased in treated ones. Again, interindividual variations were considerable (Table 2)
.
Perfusion (i.e., blood flow normalized to tissue volume) decreased in central regions of treated tumors (Fig. 9)
during the first week of therapy down to almost zero (Table 2
; Fig. 4B
), but it increased in central parts of untreated ones (Fig. 4D)
.
|
V/V; k = -0.69, P < 0.05). Accordingly, the tumor in Fig. 5
showed a marked reduction in volume after the second week of therapy (Fig. 5C)
and had already been poorly vascularized the week before (Fig. 5B)
. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Previous studies (34)
with the HaCaT-ras model have shown that suppression of ongoing angiogenesis and concomitant inhibition of tumor invasion can be achieved by a blocking antibody to VEGF receptor 2 (DC101). In our study, we used IBS to observe functional blood flow parameters and its changes under therapy in vivo. Tumor blood flow and the perfusion of the central tumor parts decreased in treated tumors as early as 1 week after the start of the treatment, whereas it increased in controls. Thus, IBS enabled us to observe tumor response on antiangiogenic therapy early and independently of a potential reduction of tumor size (Fig. 5)
by visualizing a reduction of existing functional vasculature under DC101 therapy.
To validate the IBS measurements, we compared them with histological vessel density, a well-established method to determine tumor vascularization. In general, it is difficult to compare histological slices with noninvasively obtained images. Effort was made to match the slices as well as possible. Comparison of parameters derived from IBS with the histological findings revealed a significant correlation between vessel density with the IBS parameter blood volume density but not with other parameters. Because the calculated blood volume is a nondynamic parameter, it was expected to be the closest corresponding quantity to the static measurement of vessel density. These parameters include both the macrovascularity and the microvascularity. However, because only the vessel wall but not the lumen contributes to the signal in immunofluorescence images, blood volume in larger vessels will not be adequately reflected by histology. Furthermore, vessel density does not discriminate between perfused and nonperfused vessels, and CD31-positive structures found in necrotic areas may reflect nonfunctional vessels, which are not yet disintegrated, thereby giving an insufficient picture of the actual tumor perfusion. The results showed an improved linear correlation of vessel density and calculated blood volume density when ROIs covering only the central tumor parts were included. This emphasizes the good congruence of histology and intermittent sonography, which, however, may be less appropriate at the periphery of tumors, where US could not discriminate between tumor vascularization and adjacent s.c. vessels. The other derived vascularization parameters were dynamic ones and, thus by nature, not expected to correlate with histological findings. Nevertheless, the excellent correlation between histological vessel density and blood volume clearly validate IBS as a useful and sensitive method for monitoring tumor perfusion.
Furthermore, other than conventional US methods (13, 14, 15 , 19) , intermittent sonography can visualize capillary blood flow, thereby allowing the monitoring of the highly relevant microvessel compartment during angiogenesis. In our opinion, the lack of capillary detection might be the reason why in conventional color or power Doppler studies the derived vascularization parameters of Refs. 16 , 18, 19, 20 did often not correlate with histological blood vessel counts. Thus, the IBS method developed and validated by us is clearly better suited than conventional US methods to monitor therapies which mainly affect microvascularity.
Because the number of the examined tumors was relatively low, general predictions of therapeutic effects must be discussed with caution. It is very likely that an initial high tumor blood flow and a strong reduction of tumor perfusion under therapy are correlated with the subsequent DC101-related tumor size reduction. However, previous studies (11
, 31)
had shown that perfusion of the HaCaT-ras tumors was inversely correlated with their volume, probably because of increasing necrotic areas that evolve as the tumors grow. As a consequence, these studies demonstrated a decrease in perfusion of untreated tumors during the later time points, a result that we could confirm in our study, where blood flow remained high but tumor volume additionally increased (Figs. 3
and 8
). Thus, blood flow and perfusion of HaCaT-ras tumors showed an inverse correlation with the tumors growth. The observation that the calculated perfusion is mainly reduced in the central parts of treated tumors is in line with recently performed studies of monitoring antiangiogenic therapy (DC101) with dynamic MRI.5
Furthermore, in both studies, treated tumors sometimes showed a higher global perfusion value than controls because of the strong decrease in tumor volume under treatment and thus of tumor mass, to which perfusion is normalized. This complies with our suggestion that only at early time points the degree of perfusion could be of prognostic relevance for any later tumor response. Finally, these findings suggest that vascularization parameters that are normalized to tumor size, as vessel density and perfusion, may not be a general indicator of efficacy of antiangiogenic treatment (36)
because they are also influenced by shrinkage of tumor size. More detailed studies, in particular, using early examination points after start of therapy, have to clarify whether the vascularization parameters derived with intermittent sonography will allow an early assessment of tumor response and may serve as prognostic parameters for therapeutic efficacy.
In summary, we have shown that intermittent sonography is a highly sensitive and promising method to measure several valuable perfusion parameters during early stages of antiangiogenic therapy, which are superior to static measurements such as vessel density or to conventional US parameters. Thus, using this novel technique will further our understanding of early alterations in tumor angiogenesis and allow a sensitive and quantitative noninvasive assessment of therapy-related changes of tumor vascularization.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Martin Krix, German Cancer Research Center Heidelberg, INF 280, D-69120 Heidelberg, Germany. Phone: 49-6221-422493; Fax: 49-6221-422462; E-mail: m.krix{at}dkfz.de
4 The abbreviations used are: MRI, magnetic resonance imaging; IBS intermittent bolus-contrast sonography; n.a.u., normalized arbitrary unit; ROI region of interest; US, ultrasound; VEGF vascular endothelial growth factor. 
5 F. Kiessling, N. Farhan, M. Lichy, S. Vossler, M. Heilmann, M. Krix, P. Bohlen, D. W. Miller, M. M. Mueller, H-U. Kauczor, N. E. Fusenig, and S. Delorme. Dynamic magnetic resonance imaging rapidly indicates vessel regression caused by VEGF-receptor II blockade with DC101, unpublished results.
Received 6/ 2/03. Revised 8/ 1/03. Accepted 9/16/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Palmowski, J. Huppert, G. Ladewig, P. Hauff, M. Reinhardt, M. M. Mueller, E. C. Woenne, J. W. Jenne, M. Maurer, G. W. Kauffmann, et al. Molecular profiling of angiogenesis with targeted ultrasound imaging: early assessment of antiangiogenic therapy effects Mol. Cancer Ther., January 1, 2008; 7(1): 101 - 109. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-A. Weber, H. Krakowski-Roosen, S. Delorme, H. Renk, M. Krix, J. Millies, R. Kinscherf, A. Kunkele, H.-U. Kauczor, and W. Hildebrandt Relationship of skeletal muscle perfusion measured by contrast-enhanced ultrasonography to histologic microvascular density. J. Ultrasound Med., May 1, 2006; 25(5): 583 - 591. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. E. Yankeelov, K. J. Niermann, J. Huamani, D. W. Kim, C. C. Quarles, A. C. Fleischer, D. E. Hallahan, R. R. Price, and J. C. Gore Correlation Between Estimates of Tumor Perfusion From Microbubble Contrast-Enhanced Sonography and Dynamic Contrast-Enhanced Magnetic Resonance Imaging J. Ultrasound Med., April 1, 2006; 25(4): 487 - 497. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-A. Weber, M. Krix, U. Jappe, H. B. Huttner, M. Hartmann, U. Meyding-Lamade, M. Essig, C. Fiehn, H.-U. Kauczor, and S. Delorme Pathologic Skeletal Muscle Perfusion in Patients with Myositis: Detection with Quantitative Contrast-enhanced US--Initial Results Radiology, December 21, 2005; (2005) 2382041822. [Abstract] [Full Text]
|
|
|
|
|
|
M. Krix, M.-A. Weber, H. Krakowski-Roosen, H. B. Huttner, S. Delorme, H.-U. Kauczor, and W. Hildebrandt Assessment of Skeletal Muscle Perfusion Using Contrast-Enhanced Ultrasonography J. Ultrasound Med., April 1, 2005; 24(4): 431 - 441. [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
(refilling time after local destruction of the US contrast agent) provides the calculation of perfusion parameters A and ß, which describe tumor blood volume and blood flow velocity. Straight line: nonlinear least square fitting of the measurement (
).
) is visible versus an increase of tumor volume in untreated tumors (
). Straight lines indicate median values and dotted lines the remove of tumors at the last examination time, respectively.
