ORIGINAL ARTICLE

Focusing on delayed clearance for identifying small-sizedmetastatic lung tumors using synchrotron radiation angiographywith a highly sensitive receiver

Hiromichi Ito • Shonosuke Matsushita •

Kazuyuki Hyodo • Hideo Tsurushima •

Yukio Sato • Yuzuru Sakakibara

Received: 3 April 2014 / Accepted: 19 May 2014

� The Japanese Association for Thoracic Surgery 2014

Abstract

Objectives To detect metastatic lung tumors of less than

1 mm in size by focusing on the clearance of contrast

material using synchrotron radiation (SR) angiography

characterized by high spatial resolution and high-sensitiv-

ity receiver.

Methods C6 cells, derived from rat glioma cells, were

injected to the rat tail vein. Two weeks after injection, the rats

underwent SR angiography using a high-gain avalanche

rushing amorphous photoconductor (HARP) receiver of

extra-high sensitivity with high contrast resolution. The

256-grayscale value was employed in the analysis of images.

Results 19 nodules were identified in images. The tumors

were confirmed histopathologically. The average tumor

size was 621 ± 193 lm. The clearance curve of the den-

sities was expressed as a logarithm function. Tumors

showed delayed clearance of contrast material, taking up to

28 s, compared with arteries, which cleared rapidly at 8 s.

In 256 grayscale, the distance was 50. This gap in density

clearance made it possible to identify tumors.

Conclusions SR angiography with a HARP receiver

provides high sensitivity and spatial resolution and makes

it possible to diagnose metastatic lung tumors of less than

1 mm in size by focusing on differences in the clearance

times of contrast material.

Keywords Metastatic lung tumor � Synchrotron

radiation � Angiography � Abnormal permeability � Delayed

enhancement

Introduction

The lung is a target of metastasis from various tumors. The

treatment has mainly consisted of systemic chemotherapy.

It would be important to be able to detect smaller meta-

static tumors with information on vascularity, since till now

it has been difficult to diagnose small tumors of less than

1 mm by radiographic image in vivo. This would be

Presented at the 65th Annual Scientific Meeting of the Japanese

Association for Thoracic Surgery.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s11748-014-0430-x) contains supplementarymaterial, which is available to authorized users.

H. Ito � S. Matsushita (&) � Y. Sato � Y. Sakakibara

Department of Thoracic and Cardiovascular Surgery, Graduate

School of Comprehensive Human Science, University of

Tsukuba, Tsukuba 305-8575, Japan

e-mail: gag02470@nifty.ne.jp

H. Ito

Department of Surgery, Teikyo University School of Medicine,

Itabashi-ku, Tokyo, Japan

e-mail: hiromi396@aol.com

S. Matsushita

Department of Health, Faculty of Health Science, National

University Corporation Tsukuba University of Technology,

4-12-7 Kasuga, Ibaraki, Tsukuba 305-8521, Japan

K. Hyodo

High Energy Accelerator Research Organization: KEK,

Tsukuba, Japan

H. Tsurushima

Department of Neurosurgery, Graduate School of

Comprehensive Human Science, University of Tsukuba,

Tsukuba, Japan

123

Gen Thorac Cardiovasc Surg

DOI 10.1007/s11748-014-0430-x

especially important for patients showing elevation of

tumor markers in advanced stages of cancer [1].

PET–CT cannot be used to clearly diagnose tumorous

shadows of around 5 mm, which is equivalent to human

acinar size, mostly because of limitations in spatial resolution

[2–4]. And although there has been improvement in high-

resolution CT, small tumors of less than 1 mm cannot be

diagnosed even if visualized as dots [5]. Usually, metastatic

lung tumors are rich in vessel size variety due to tumor vas-

culogenesis [6, 7]. A diagnostic concept for metastasis

focusing on hyperpermeability in CT was proposed in 1997

[8]. CT scans, possibly in connection with the neovascularity

in tiny tumor deposits, may be used to predict the future

appearance of macroscopic hepatic metastases [8]. These are

examples of how tumor-related angiogenesis can improve the

detection of small malignant foci. However, it is believed that

micrometastases have never been visible in imaging at any

time during the twentieth century [9, 10].

In angiography, since tumor vessels indicate special

properties such as abnormal permeability [11], the detection

of a wide spectrum of contrast density may be of help in

identifying smaller nodules as tumors [10]. In this context

both high spatial resolution and high sensitivity are required.

Synchrotron radiation (SR) is characterized by a uni-

form white light with high photon density and high direc-

tivity which can be used for obtaining high-resolution

images for medical applications [12]. Visualization of

small vessels and microstructure in the lung by synchrotron

radiation has already been investigated experimentally

[13–15]. In addition, a high-sensitivity receiver named the

high-gain avalanche rushing amorphous photoconductor

(HARP) receiver, developed by NHK Science and Tech-

nology Research Laboratories (Tokyo, Japan) can theo-

retically be considered to be approximately 100 times more

sensitive than conventional CCD cameras [16, 17]. Over

the past 10 years we have been working to improve the

effectiveness of high-resolution angiography using syn-

chrotron radiation in vivo [18], and it is now possible to

visualize down to a coronary artery spasm of about 50 lm

with the use of a HARP receiver (unpublished data).

In this experiment, the subject is to detect metastatic

lung tumors of less than 1 mm in size by focusing on the

clearance of contrast material. We hypothesized that syn-

chrotron radiation angiography with a highly sensitive

receiver would make it possible.

Methods

Animal model

The Committee on Animal Research at the University of

Tsukuba approved the experimental protocols. The rats

were cared for in accordance with the Guiding Principles

for the Care and Use of Animals based on the Helsinki

Declaration of 1964. The experiment is also in accordance

with standard guidelines by the Science Council of Japan.

Male Wistar rats (6 weeks, 200 g, Charles River, Japan,

n = 4) were employed to detect metastatic lung tumor (rat

no. #1–#4).

Tumor cells

C6 cells, derived from rat glioma cells (RIKEN, Japan)

were incubated in 1640 medium containing 10 % fetal

bovine serum. Fifty microliters of phosphate buffer saline

(PBS) containing 105 of C6 cells were injected to the tail

vein to create the lung metastasis model [19].

Synchrotron radiation angiography

Synchrotron radiation

When high-energy electrons which go straight at near the

speed of light are bent with a strong magnetic field,

electromagnetic waves are discharged by the tangential

projection of the trace [20]. These electromagnetic waves

are called synchrotron radiation (SR) [21, 22]. SR angi-

ography (SRA) was performed at the Photon Factory-

Accelerator Ring (PF-AR) at the High Energy Accelerator

Research Organization in Tsukuba, Japan. Our SR was

obtained from a 6.5-GeV electron beam. SR was con-

verted to monochromatic X-rays by reflecting at 13� on a

silicon crystal. The energy of the monochromatic X-rays

was 33.3 keV with high photon density (10000 times that

of conventional X-rays) and a straight beam, leading to

high spatial resolution [21]. The X-rays were converted to

visible light on a fluorescent screen made of cesium

iodide.

The HARP receiver

In the HARP receiver, amorphous selenium is used for the

photoelectric conversion film which converts light into an

electrical signal [17]. When light is incident on the film,

hole–electron pairs are generated, and accelerated in high

electric fields. These are ionized by colliding and form new

hole–electron pairs. The collision and ionization process is

repeated, and the electric signal is multiplied. This is called

the avalanche multiplication phenomenon [17]. The signal

electric current output grows extremely large as a result of

this multiplication phenomenon. Because television camera

technology is used, the exposure time was fixed at 30 ms/

frame [21]. In this study, optimization was achieved in

lens-coupling between the fluorescent screen and the

HARP receiver [16].

Gen Thorac Cardiovasc Surg

123

Experimental protocol

Angiographic procedure Two weeks after tumor cell injec-

tion, the rats underwent SRA. Anesthesia was performed by

inhalation of diethyl ether, followed by intraperitoneal

injection of pentobarbital. The right jugular vein was punc-

tured by a 24-gauge needle and was cannulated with a plastic

catheter sheath into the junction to the superior vena cava.

A contrast material (32 % non-ionic iodine) was admin-

istered into the right jugular vein by a programmed injector

at 1 ml/s for 2 s. Consecutive images were recorded, and

each frame was converted into still images for analysis.

CT scanner CT scan was performed using an animal CT

scanner accessible to rats (LCT100Lite�, ALOCA, Japan).

Evaluation

Analysis of images

Smaller nodules which were candidates of detected

metastasis were marked using alphabetical codes. The

256-grayscale value was employed using the graphic

software Image J (NIH, USA). The grayscale value of each

nodule represented the difference of X-ray absorption of

iodine. High grayscale values indicated a high value of

photon absorption by iodine. The grayscale time course in

suspected metastatic nodules was measured. Grayscale

values of the subpleural peripheral lung and the main

pulmonary artery (PA) in the same right lower lobe were

also measured for comparison.

Histopathological analysis

The lungs were resected immediately after SRA and fixed

with an intratracheal injection of 10 % formalin solution

under a constant pressure of 20 cmH2O. Metastatic tumors

were also pathologically detected by H&E staining of the

number of sections, which came to a total of 60 slides of

2 cm in length from the right lung, with slices made every

330 lm.

Results

Angiographic findings

Out of all the rats, there were 19 nodules which were

suspected of being metastatic lung tumors in the SR

angiography. Number of tumor in each rat (rat no. #1–#4)

was (4, 7, 4, 4, respectively) in SR angiography. A case of

SR angiography of rat (rat no. #2) is shown in the anima-

tion (Online Resource 1).

Clearance curve of contrast material densities

The clearance of averaged contrast material density in all 19

nodules, as well as that of PA and the peripheral lung field

density of all the rats are shown as a plotted graph (Fig. 1).

The peak densities of the nodules and the PA were the same.

After the peak density of PA at 1–2 s on the X axis, the PA

density began to decrease rapidly for about 10 s, down to the

peripheral lung level followed by a plateau. By contrast in

nodules, after a rapid decrease of density in the initial 8 s from

the peak density, nodule densities showed a slow decrease

continuing up to approximately 28 s. Dots in each group are

expressed as a logarithm function with a correlation coeffi-

cient (p \ 0.001, each). The difference in density became

larger at around 8 s. Clear identification was possible when

the larger distances in contrast between the nodules and PA or

peripheral lung remained for from 8 to 28 s (Fig. 1).

Typical SR angiographic still images of metastatic lung

tumors

Looking at one of the typical SR angiographic still images

of the metastatic tumors, it can be seen that small multiple

lung metastasis became clearer compared with the back-

ground in the late phases of 15 and 25 s (Fig. 2c, d) than in

the early phase (Fig. 2b). Most of the nodules were not

visible before injection of the contrast material (Fig. 2a).

Histopathologically, an example of a clear metastatic lung

tumor is shown in Fig. 3. The size of this tumor was

Fig. 1 Time course of the densities of all 19 nodules from 4 rats

compared with the PA and peripheral lung field after injection of

contrast material. The time course of the average density of all 19

nodules, PA and peripheral lung field from four rats were significantly

comparable to the logarithm function curve. The coefficients of

determination (R2 value) were 0.91 (nodule), 0.91 (PA), and 0.76

(peripheral lung field), respectively. Nodule densities showed a

relatively slow decrease from 8 to 28 s, whereas PA and peripheral

lung fields continued to decrease rapidly compared with nodules. This

period is optimal timing for diagnosing tumors by utilizing larger

distances of density between tumors and backgrounds

Gen Thorac Cardiovasc Surg

123

Fig. 2 The pulmonary SR

angiography. Smaller nodules as

candidates of metastasis

detected were marked with

alphabetical codes. a Before

injection. Only two equivocal

nodules were seen (marked by

J, L). b This still image was

captured one second after

injection. Tumorous shadows

were not clearly identified. c,

d Several small-sized tumors

were identified clearly, due to

delayed clearance of density

from 8 up to 28 s after injection

of contrast material. These

tumors were approximately

500 lm in size (encircled and

marked with letters)

Fig. 3 Histopathological

findings of the identified

metastatic tumor nodules.

a (H&E, 9100) A well-defined

round shaped cell-cluster was

identified in the right lower

lobe. b, c (H&E, 9400) The

features of the cell-cluster,

which contained concentrated

highly dysplastic cells whose

nuclei were large and showed

nuclear atypia, were compatible

with metastatic lung tumor. b is

magnification of part of a (black

thick arrow). c is the other

tumor in the same lung which is

also containing highly

dysplastic cells. d (H&E,

91000) Furthermore, typical

nuclear division (encircled)

found in the other tumor of

the same lung was shown

Gen Thorac Cardiovasc Surg

123

450 lm at 100-fold magnification, corresponding to the

angiographic image of the tumor (Fig. 3a).

Tumor size

Tumor size ranged from 400 to 1000 lm which was

derived from angiographic image (all of the 19 tumorous

nodules in SRA), and the average size of the total of 19

nodules was 621 ± 193 lm. The average of pathological

tumor size was 235 ± 146 lm. Even in smaller sizes of

around 500 lm, tumors were clearly visualized (Fig. 2).

Histopathological findings

From histological examination by a pathologist who is not

involved in this study, we could obtain 17 tumorous lesions

which consisted of atypical cell clumps with enlarged or

deformed nuclei. The number of tumor in each rat (rat no.

#1–#4) was (6, 3, 4, 4, respectively) in histopathology. We

selected the most typical tumor to demonstrate from lung

of a rat (no. #1) as Fig. 3. A well-defined cell-cluster was

identified in the left lower lobe (Fig. 3a). The cell-cluster

contained concentrated highly dysplastic cells with vascu-

lature (Fig. 3b, c). The size of the cell-cluster was 450 lm

at 100-fold magnification (Figs. 2d, 3a).

The cell-cluster was compatible with metastatic lung

tumor of cultured glioma cells, because nuclei were large

and showed nuclear atypia (Fig. 3b, c) and partial nuclear

division at 1000 times magnification (Fig. 3d).

Conventional plane CT scan

CT scan was taken just before SR angiography. However,

any tumorous shadows could not be distinguished other

than pulmonary vessels (Fig. 4).

Discussion

In our SRA system, 19 tumorous shadows were captured

2 weeks after injection. This is the first report known to us

in which small pulmonary metastatic tumors of approxi-

mately 500 lm in size have been diagnosed using radio-

graphic imaging in vivo. There was some discrepancy

between angiographic image size and histological size, on

the other hand, none was found more than 1 mm in either

SRA or histology of specimens. The latter was probably

due to short time duration for tumor growth in this study.

Accountable reasons for this discrepancy might be as fol-

lows: (1) cut surface in histology was not always maximal

in diameter compared with 2D in angiographic image; (2)

alveolar space tended to collapse in the specimen, whereas

alveolar space was inflated in living rats; (3) in tumor

angiography, because enhanced permeability of tumor

vessels can affect contrast material in vessels to move

across the vessel walls, the visible border of tumorous

shadow in SR angiography might become larger than ori-

ginal tumor size. However, size of the tumor will be dis-

cussed from radiological findings in clinical settings in

future. Apart from the tumor size, the discrepancy of

number of tumor in each rat was little. The discrepancy of

number of tumors might be due to technical limitation

especially from treatment of specimen such as embedding

and thin slice in histological examination.

The use of SRA with a highly sensitive receiver made it

possible to visualize such smaller metastatic lung tumors,

and thus our hypothesis was proven. Focusing on the

delayed clearance of contrast material was extremely use-

ful for radiographically identifying possible small meta-

static lung tumors which were later confirmed as tumors by

histopathological examination (Figs. 2a–d, 3a–d). These

nodules could not be identified with conventional CT scans

Fig. 4 Transverse section of rat thorax by thin slice CT scan. The

thickness of each slice was 1 mm. The size of pulmonary arteries

shown in the lower image was around 2–3 mm. The suspicious

nodules, which were less than 1 mm, could not be identified

Gen Thorac Cardiovasc Surg

123

(Fig. 4). This may be related to the spatial resolution lim-

itations of CT scan.

The reasons for the good correspondence between the

nodules in the images and the histological specimens can

be attributed to the following radiographic properties: (1)

spatial resolution, (2) density resolution, and (3) time res-

olution. The spatial resolution of this system is 16 lm/

pixel, which makes it possible to identify arterioles of

50 lm in diameter [18]. High-density resolution is mainly

derived from the HARP receiver, which can amplify the

dynamic range of gray scale by around several hundred

times compared with conventional CCD cameras [16, 17].

The high-density resolution might contribute to identifying

nodules overlapping with other pulmonary structures. For

example, nodules could be identified despite overlapping

with the ribs in Fig. 2c (G, J, L, M, N). The time resolution

was utilized to compare the average clearance curve of

each density in pulmonary structure. The densities of

contrast material in tumorous nodules were more pro-

longed against a downward shift than those in PA and

background peripheral lung tissue (Figs. 1, 2). The

enhanced tumorous nodules seem to reflect a prolonged

retention of contrast material in neoplasm.

Angiogenesis of tumors plays an important role in tumor

development and tumor metastasis and is regarded as a

hallmark of cancer [6, 23]. In general, tumor angiogenesis

characterizes: (1) abnormal vascular density, (2) weakness,

(3) increased permeability, and iv) serpiginosum [6, 23].

Among these, it is notable that metastatic lung tumor is

accompanied by immature angiogenesis demonstrating

abnormal vascular permeability [6, 7, 24]. Generally,

tumor angiogenesis is initiated by the onset of severe

central hypoxia accompanied by tumor growth. It is

reported that the oxygen contents drop to zero when

200 lm is taken from capillaries in tumors [25]. Conse-

quently, HIF-1a and VEGF are upregulated [26]. Due to

cytogenetic abnormalities of the tumor vessel endothelium,

increased permeability and protein leakage emerge in

relation to increased VEGF concentrations [9, 27].

In this study, delayed clearance of density in nodules

seemed to correspond to the topical dynamics of the

contrast material. Indeed, it is reported that contrast

enhancement and increased signal on T2-weighted MR

images are the imaging correlates of the increased

microvascular density and vascular permeability that

occurs with tumor angiogenesis [9]. Because tumor ves-

sels are more prone to leaking substances with large

molecules than are vessels in normal tissues [28], mac-

romolecular contrast material may enhance more specifi-

cally for tumors, resulting in crossing tumor vessels but

not normal vessels [10]. Recently, methods of detecting

leakage of macromolecular contrast material utilizing

MRI or CT scans have been reported [29, 30]. In our

radiographic system, it is suspected that the leakage and

salvage of contrast material is visualized in small-sized

metastatic tumor.

Clinical implications of this system can be considered

for high-risk cancer patients such as lung metastasis sus-

pected because of slight elevation of tumor markers

without apparent symptoms, even in use of high-resolu-

tion CT scan. As a consequence of findings on highly

suspicious, metastatic lung tumors, early initiation of

optimal therapies such as chemotherapy or molecular-

targeted drugs could be selected after pathological con-

firmation by lung biopsy with minimal resection, and

could then be combined with radiation therapy, or mini-

mal invasive resection.

Considering clinical applications of this system, the

following improvements may be required: (1) downsizing

the SR generator, and (2) reducing radiation exposure. For

downsizing the SR generator, inverse Compton scattering

could possibly be used for the generation of synchrotron

radiation. For a future light source, small-sized SR gener-

ators (comparable to a large office desk) are also under

investigation. Radiation exposure with synchrotron radia-

tion has been high [13]. Our previous study demonstrated

that the limitation of the continuous exposure time for

patient safety was 105.8 s for 60 mSV [16]. Furthermore,

in response to this problem, intermittent exposure may be

feasible for diagnosing metastatic lung tumor. For exam-

ple, in our study just four still image frames at 10, 15, 20

and 25 s after injection were sufficient for diagnosing

metastatic tumor (Fig. 2). Thus, the radiation dose from

four still image frames should be minimized down to

7.60 9 10-2 mSV at 4 frames [16], which is expected to

be around a hundredth part of the radiation exposure of

conventional CT scan.

Conclusion

Analysis of the delayed clearance of density with the use of

SRA with HARP made it possible to diagnose metastatic

lung tumors of less than 1 mm in diameter in rats. This

system may provide beneficial potentials, such as early

diagnosis to facilitate early initiation of optimal therapies.

Limitation

In the future, the topical hemodynamics of contrast mate-

rial across the tumor vessels needs to be thoroughly

examined in regard to abnormal permeability.

Acknowledgments The authors thank Mr. Kenkichi Tanioka of

Tokyo Denki University, Mr. Misao Kubota, Mr. Kazunori Miyakawa

of NHK Science and Technical Research Laboratory and Mr. Akira

Gen Thorac Cardiovasc Surg

123

Kobayashi of Hamamatsu Photonics Corporation for their technical

assistance. The authors are grateful to Mr. Avi Landau for his lan-

guage consultation. This study was supported in part by a Grant-in-

Aid for Scientific Research from the Japan Society for the Promotion

of Science (Grant C-20591471 and Grant C-23592053).

Conflict of interest The authors have declared that no conflict of

interest exists.

References

1. Smith RA, Cokkinides V. American cancer society guidelines for

the detection of cancer. CA Cancer. 2002;52:8–22.

2. Deppermann KM. Lung cancer screening-where we are in 2004.

Lung Cancer. 2004;45:S39–42.

3. McDonald DM, Choyke PL. Imaging of angiogenesis: from

microscope to clinic. Nat Med. 2003;9:713–25.

4. Koh DM, Cook GJ, Husband JE. New horizons in oncologic

imaging. N Engl J Med. 2003;348:2487–8.

5. Battista G, Sassi C, Zompatori M, Palmarini D, Canini R.

Ground-glass opacity: interpretation of high resolution CT find-

ings. Radiol Med. 2003;106:425–44.

6. Hannahan D, Wein berg RA. The hallmarks of cancer. Cell.

2000;100:57–70.

7. Folkman J. Angiogenesis in cancer, vascular, rheumatoid and

other disease. Nat Med. 1995;1:27–31.

8. Leggett DAC, Kelley BB, Bunce IH, Miles KA. Colorectal

cancer: diagnostic potential of CT measurements of hepatic

perfusion and implications for contrast enhancement protocols.

Radiology. 1997;205:716–20.

9. Kuszyk BS, Corl FM, Franano FN, Bluemke DA, Hofmann LV,

Fortman BJ, et al. Tumor transport physiology: implications for

imaging and imaging-guided therapy. AJR. 2001;177:747–53.

10. Gossmann A, Okuhata Y, Shames DM, Helbich TH, Roberts TP,

Wendland MF, et al. Prostate cancer tumor grade differentiation

with dynamic contrast-enhanced MR imaging in the rat: com-

parison of macromolecular and small-molecular contrast media—

preliminary experience. Radiology. 1999;213:265–72.

11. Kuszyk BS, Boitnott JK, Choti MA, Bluemke DA, Sheth S,

Magee CA, et al. Local tumor recurrence following hepatic

cryoablation: radiologic–histologic correlation in an animal

model. Radiology. 2000;217:477–86.

12. Schwenke DO, Pearson JT, Umetani K, Kangawa K, Shirai M.

Imaging of the pulmonary circulation in the closed-chest rat using

synchrotron radiation microangiography. J Appl Physiol.

2007;102:787–93.

13. Ikura H, Shimizu K, Ikezoe J, Nagareda T, Yagi N. In vitro

evaluation of normal and abnormal lungs with ultra-high-reso-

lution CT. J Thorac Imaging. 2004;19:8–15.

14. Bayat S, Porra L, Suhonen H, Nemoz C, Suortti P, Sovijarvi AR.

Differences in the time course of proximal and distal airway

response to inhaled histamine studied by synchrotron radiation

CT. J Appl Physiol. 2006;100:1964–73.

15. Kono M, Ohbayashi C, Yamasaki K, Ohno Y, Adachi S, Su-

gimura K, et al. Refraction imaging and histologic correlation in

excised tissue from a normal human lung: preliminary report.

Acad Radiol. 2001;8:898–902.

16. Konishi T, Matsushita S, Hyodo K, Sato F, Hiramatsu Y, Sak-

akibara Y. Examination of a new angiographic system that uti-

lizes a highly sensitive receiver and synchrotron radiation for

reducing the dose of contrast medium. Am J Roentgenol.

2011;197:W508–13.

17. Tanioka K, Matsubara T, Ohkawa Y, Miyakawa K, Suzuki S,

Takahata T, et al. Ultra-high-sensitivity new super-HARP pickup

tube and its camera. IEICE Trans Electron. 2003;E86-

C(9):1790–5.

18. Matsushita S, Hyodo K, Imazuru T, Tokunaga C, Sato F, E-

nomoto Y, et al. The minimum coronary artery diameter in which

coronary spasm can be identified by synchrotron radiation coro-

nary angiography. Eur J Radiol. 2008;68:S84–8.

19. Mandybur TI, Liwnicz BH, Wechsler W. Disseminated (meta-

static) tumors in nude mice produced by intravenous injection of

cells of human and nonhuman neurogenic tumor lines. Acta

Neuropathol. 1984;63:203–9.

20. Arfelli F. Synchrotron light and imaging systems for medical

radiology. Nucl Instrum Methods Phys Res A. 2000;454:11–25.

21. Miyakawa K, Ohkawa Y, Matsubara T, Kikuchi K, Suzuki S,

Tanioka K, et al. Development of FOP-HARP imaging device.

Sensors, cameras, and systems for industrial/scientific applica-

tions XI SPIE-IS&T. 2010; 7536:753604-1-8.

22. Meuli R, Hwu Y, Je JH, Margaritondo G. Synchrotron radiation

in radiology: radiology techniques based on synchrotron sources.

Eur Radiol. 2004;14:1550–60.

23. Carmeliet P, Jain RK. Angiogenesis in cancer and other disease.

Nature. 2000;407:249–57.

24. Fink C, Kiessling F, Bock M, Lichy MP, Misselwitz B, Peschke

P, et al. High-resolution three-dimensional MR angiography of

rodent tumors. J Magn Reson Imaging. 2003;18:59–65.

25. Helmlinger G, Yuan F, Dellian M, Jain RK. Interstitial pH and

pO2 gradients in solid tumors in vivo: high-resolution measure-

ments reveal a lack of correlation. Nat Med. 1997;3:177–82.

26. Furuya M, Yonemitsu Y. Cancer neovascularization and proin-

flammatory microenvironments. Curr Cancer Drug Targets.

2008;8:253–65.

27. Hida K, Hida Y, Amin DN, Flint AF, Panigrahy D, Morton CC,

et al. Tumor-associated endothelial cells with cytogenetic

abnormalities. Cancer Res. 2004;64:8249–55.

28. Gerlowski LE, Jain RK. Microvascular permeability of normal

and neoplastic tissues. Microvasc Res. 1986;31:288–305.

29. Roberts HC, Roberts TP, Brasch RC, Dillon WP. Quantitative

measurement of microvascular permeability in human brain

tumors achieved using dynamic contrast-enhanced MR imaging:

correlation with histologic grade. AJNR Am J Neuroradiol.

2000;21:891–9.

30. Cuenod CA, Fournier L, Balvay D, Guinebretiere JM. Tumor

angiogenesis: pathophysiology and implications for contrast-

enhanced MRI and CT assessment. Abdom Imaging.

2006;31:188–93.

Gen Thorac Cardiovasc Surg

123