Breast CancerResearchand Treatment33: 269-273,1995. © 1995KluwerAcademic Publishers.Printedin the Netherlands. Brief communication

Relationship between staining by digital subtraction angiography and vascularity in breast cancer: Analysis of the time-density curve and the results of staining for factor VIII-related antigen/von Willebrand factor Osamu Watanabe, Shunsuke Haga, Tadao Shimizu, Hiroshi Imamura, Koji Kobayashi, Jun Kinoshita, Tetsuro Kajiwara 1and Mariko Fujibayashi 2 l Department of Surgery, Tokyo Women's Medical College Daini Hospital, Tokyo, Japan 2 Department of Pathology, Tokyo Women's Medical College Daini Hospital, Tokyo, Japan

Key words: DSA for breast cancer, factor VIII-related antigen/von Willebrand factor, angiogenesis, vascularity

Summary We performed intravenous digital subtraction angiography (IV-DSA) in 31 patients with primary invasive breast cancer, and calculated the maximum density of tumor staining from the time-density curve obtained. Specimens resected from the same patients were stained for factor VIII-related antigen/yon Willebrand factor (vWF), and the amount of microvessels was calculated by image processing. A significant correlation (correlation coefficient: 0.85) was found between the maximum density of tumor staining and the vascularity determined by staining for vWF, indicating that the maximum density indirectly reflects the vascularity in the tumor.

Introduction Digital subtraction angiography (DSA) is widely applied to patients treated in general surgical fields, including those with breast diseases [1]. In particular, intravenous DSA (IV-DSA) is minimally invasive, and useful for determining whether a particular breast disease is benign or malignant [1]. Since DSA provides digital data of images, analysis of the data can disclose new aspects of the information. In the present study, we quantified the time-density curve (TDC) obtained in staining of breast cancers by IV-DSA, and calculated the maximum density from the curve [2]. Tumors resected from the same patients were then stained with antibody against factor VIII-related antigen (von Willebrand factor,

vWF), and the amount of microvessels was determined by image processing to investigate the relationship between staining by DSA and vascularity.

Patients and methods Patients

The subjects of the present study were 31 patients with primary invasive ductal carcinomas measuring 5 cm or less and treated in Tokyo Women's Medical College Daini Hospital between July 1987 and January 1991.

A ddressfor offprints: O. Watanabe, Department of Surgery,TokyoWomen'sMedicalCollegeDaini Hospital,2-1-6Nishi-okuArakawaku, Tokyo,Japan

270

0 Watanabe et aL Pixel

8.5972

5

'

1

'

'

'

I

6

'

'

'

I 1 !

11

|

v

~

I

'

16

'

!

'

I

i

21

,

i

!

I

'

i

26

i

i

30 sec

Time Density Curve (TDC) ROI on the tumor stain Fig. l. Tumor stain and the time-density curve.

IV-DSA IV-DSA was carried out using a Shimazu digital subtraction angiography system, DAR1200, which has a matrix of 512 x 512 pixels. For IV-DSA, no catheter was inserted, but a 16-gauge elastic catheter needle was retained in the cubital vein on the contralateral side. A bistratal solution composed of Iopamidol 40 cc and physiological saline 40 cc was infused automatically at a rate of 15-16 cc/sec. Lateral views of the breast were obtained while the patient was lying about 45 ° obliquely on the affected side. DSA images were processed at 30 frames/sec, and the continuous mode was used for image collection.

TDC The region of interest (ROI) was set in the stained areas of IV-DSA images, and the density determined every 0.5 sec was plotted on a graph (Fig. 1).

The maximum density was calculated by subtracting the background density from the maximum value of the plotted density.

Staining for vWF Two sections giving the maximum cross-sectional area of the tumor were prepared from each paraffin-embedded tissue block. The sections were stained by immuno-histochemical staining using the peroxidase anti-peroxidase method [3]. Antifactor VIII rabbit serum (DAKO, Sweden) was used as the specific antibody, and peroxidase-labeled anti-rabbit IgG (MBL, Japan) was used as the labeled antibody. The sections were then subjected to coloring with dimethylaminobenzidine HzO 2 solution and nuclear staining with methyl green. After dehydration and mounting, they were examined microscopically. Details of this procedure are shown in Table 1.

Angiography for breast cancer

271

Fig. 2. Image analysis by Olympus color image analyzer. Original view obtained by staining for vWF (left). Stained part was emphasized through processing by computer (right).

Image processing

Table1. Procedures of staining for factor VIII-related antigen Preparation of a paraffin block, sectioning Deparaffinization with xylol Treatment with 0.1% H202 - PBS solution 5 rain at room temperature Treatment with 0.1% trypsin • Tris solution 30 min at 37° C Treatment with 10% normal swine serum 20 rain at room temperature Specific antibody reaction: anti-factor VIII rabbit serum overnight at 4° C Labeled antibody reaction: peroxidase-labeled anti-rabbit IgG 1 h at 37° C Coloring (DAB • H202 solution) 5 rain at room temperature Nuclear staining (methyl green) 20 rain at room temperature Dehydration, mounting Microscopy

T h e original view o b t a i n e d by staining for v W R was p r o j e c t e d on the m o n i t o r of an O l y m p u s color image analyzer, SP-500, and the b r o w n - s t a i n e d part was identified t h r o u g h digital processing by c o m p u t e r (Fig. 2). T h e area of the stained p a r t r e c o g n i z e d by c o m p u t e r was expressed as the area per visual field, and used as the index of vascularity.

Results T h e r e was a significant correlation b e t w e e n the m a x i m u m density of staining by I V - D S A and vascularity d e t e r m i n e d by staining for vWF, showing a correlation coefficient of 0.85 (p < 0.01) (Fig. 3).

272

0 Watanabe et al.

I%1 5.0" "2

2.5-

"6 •

I--

y = - 0.14 +

0.29x

r = 0.85(P<0.01)

0-

;

;

1;

1;

The maximum density

of tumor

2; staining

Fig. 3. The area of vwF staining and the maximum density of tumor staining.

Discussion It has been reported that the growth of cancer is closely related to angiogenesis [4, 5], and the degree of angiogenesis is related to metastasis. In other words, the higher the vascularity, the higher the incidence of metastasis [4]. In this regard, we sought a simple method for determining the amount of microvessels in a tumor. We use IV-DSA currently in patients with breast cancer prior to surgery. This procedure is simple and minimally invasive, and can be completed within a short period of time. Assuming that staining by IV-DSA would reflect the vascularity of the tumor, we examined the correlation between the results of the staining and pathological features. Specifically, we quantitated the density of DSA staining, determined the amount of microvessels in the maximum cross-section of the tumor by means of vascular wall staining, and examined their correlation with each other. It is difficult to identify microvessels in a tumor by the usual hematoxylin-eosin staining. It is particularly difficult to distinguish histologically by this staining between the microvessels and lymph vessels or mammary ducts in breast cancer tissue. Factor VIII/yon Willebrand factor (vWF) (formerly called factor VIII-related antigen) is a protein involved in platelet thrombus formation and produced in vascular endothelial cells [6, 7]. Staining for vWF stains the endothelia of arterioles, capillaries, and venules, but not the lymphatic endothelium [8]. Therefore, this antigen has recently become widely used as a representative marker in demonstration of vascular endothelial cells [3, 6], and has

also been reported to be useful for determination of vascular invasion inside the tumor [9]. Because of this, we used this method for staining of the maximum cross-sections. The microvessels in the tumor were stained light brown. Weidner et al. [4] counted the density of staining and graded it. We identified the stained areas alone through digital processing by computer, and calculated the area of staining per field for use as an index of vascularity. Tumor staining by IV-DSA can be easily quantified by drawing a time-density curve for the staining. While correlation between tumor staining by IV-DSA and the pathology has not so far been documented in detail, our present study demonstrated that the density of staining represents the vascularity. In other words, the vascularity in a tumor can be determined indirectly by DSA, without complicated processes of histological staining. We have performed IV-DSA in 190 patients with breast cancer from July 1987 to date, and determined the maximum density of tumor staining in these patients. We intend to further accumulate our clinical cases and examine the relation between the vascularity of the tumor, clinicopathological factors, and prognosis.

Acknowledgements We are grateful to Dr. Takayuki Yamada and Mr. Hiroshi Murakami for excellent technical assistance.

References 1. Watt AC, Ackerman LV, Shetty PC et al.: Breast lesions: differential diagnosis using digital subtraction angiography. Radiology 159: 39-42, 1986 2. Watanabe O, Haga S, Kobayashi K, Ishibashi K, Kinoshita H, Imamura H, Iida T, Shimizu T, Kumazawa K, Kajiwara T, Yamada T: Digital subtraction angiography for breast canceranalysis of the time density curve. Jpn J Breast Cancer 7: 582586,1992 (Abstract in English) 3. Mukai K, Rosai J, Burgdorf WHC: Localization of factor VIII-related antigen in vascular endothelial cells using immunoperoxidase method. Am J Surg Pathol 4: 273-276, 1984 4. Weidner N, Semple JR Welch WR, Folkman J: Tumor angiogenesis and metastasis- correlation in invasive breast carcinoma. N Engl J Med 324: 1-8,1991

Angiography for breast cancer 5. Folkman J: What is the role of endothelial cells in angiogenesis? Lab Invest 51: 601-604, 1984 6. Sehested M, Hou-Jense K: Factor VIII related antigen as an endothelial cell marker in benign and malignant diseases. Virchow Arch [Pathol Anat] 391: 217-225, 1981 7. Ohtani H, Sasano N: Characterization of micro-vasculature in the stroma of human colorectal carcinoma: an immunoelectron microscopic study on factor VIII/von Willebrand factor. J Electron Microsc 36: 204-364, 1989

273

8. Beckstead JH, Wood GS, Fletcher V: Evidence for the origin of Kaposi's sarcoma from lymphatic endothelium. Virchows Arch [Pathol Anat] 391: 217-225, 1981 9. Kato T, Kimura T, Muraki H, Kamio T, Fujii A, Yamamoto K, Hamano K, Hirayama A: Studies of pathological factors and blood vessel invasion by using the immunoperoxidase technique for factor VIII related antigen in breast cancer. J Jpn Soc Clin Surg 51: 1397-1403, 1990 (Abstract in English)