Angiogenesis and VEGF expression in pre-invasive lesions of the human breast

Transcription

1 Journal of Pathology Published online in Wiley InterScience ( DOI: /path.1626 Original Paper Angiogenesis and VEGF expression in pre-invasive lesions of the human breast Paolo Viacava, 1 * Antonio Giuseppe Naccarato, 1 Guido Bocci, 2 Giovanni Fanelli, 1 Paolo Aretini, 1 Antonino Lonobile, 1 Giuseppe Evangelista, 3 Giancarlo Montruccoli 4 and Generoso Bevilacqua 1 1 Division of Surgical, Molecular and Ultrastructural Pathology, Department of Oncology, University of Pisa and Pisa University Hospital, Pisa, Italy 2 Division of Pharmacology and Chemotherapy, Department of Oncology, University of Pisa and Pisa University Hospital, Pisa, Italy 3 Department of Surgery, University of Pisa, Italy 4 Madre Fortunata Toniolo Hospital, Division of Gynaecology and Obstetrics, Bologna, Italy *Correspondence to: Dr Paolo Viacava, Department of Oncology, Division of Surgical, Molecular and Ultrastructural Pathology, Via Roma 57, I Pisa, Italy. pviacava@med.unipi.it Received: 1 December 2003 Revised: 16 June 2004 Accepted: 23 June 2004 Abstract Angiogenesis (as microvascular density MVD) and vascular endothelial growth factor (VEGF) expression were evaluated by immunohistochemistry in all types of human preinvasive breast lesion, un-associated with invasive carcinoma, including florid ductal hyperplasia of usual type (FDHUT, 40 cases), atypical ductal hyperplasia (ADH, 10), welldifferentiated intraductal carcinoma (WDIC, 16), intermediately differentiated intraductal carcinoma (IDIC, 25), poorly differentiated intraductal carcinoma (PDIC, 20), atypical lobular hyperplasia (ALH, 13), and lobular carcinoma in situ (LCIS, 12). Both parameters were also studied in normal glandular structures obtained from normal breasts or from breasts containing pre-invasive lesions. Increased vascularization was present in all lesion types (MVD mean values (expressed as vessel number/mm 2 ): 115 ± 8 in normal lobules, 146 ± 26 in lesions; p < 0.05) and increased with lesion severity. In ductal lesions, MVDs were significantly higher in PDIC (190 ± 65) and IDIC (167 ± 61) than in FDHUT (123 ± 40) and ADH (122 ± 47); MVD was much higher in PDIC than in WDIC (p < 0.001). WDIC showed peculiar features, with a degree of vascularization closer to hyperplasia than to the other histological types of in situ ductal cancer; this observation is in line with the hypothesis that IDIC and PDIC may originate de novo, without a mandatory transition through WDIC. LCIS was more vascularized than ALH (168 ± 50 and 125 ± 40, respectively; p < 0.05), showing MVD values similar to those of PDIC and IDIC. The vascularization of normal lobules was constant, regardless of their association with lesions. VEGF expression in normal glandular structures was lower than in lesions, with the highest levels found in ductal lesions when compared with lobular lesions. No correlation was found between VEGF expression and the degree and/or type of vascularization. Copyright 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd. Keywords: angiogenesis; VEGF; pre-invasive breast lesions; breast carcinoma in situ; breast hyperplasia Introduction Angiogenesis, the process of forming new blood vessels from the existing vascular network, is a critical event in tumour growth and metastatic spread. Quantification of angiogenesis, by microvessel density (MVD), is considered to be a valuable prognostic indicator of neoplastic aggressiveness [1 3]. The degree of vascularity in tumour tissues has been highlighted by immunohistochemistry, using a variety of antibodies. In particular, antibodies against CD34, a protein associated with endothelial cells [4], have proven to be very reliable [5 7]. There is some evidence that angiogenic activity is a marker of pre-invasive lesions, as observed in cervical dysplasia [8] and in pre-neoplastic lesions of the bladder [9]. Increased angiogenesis in the stroma surrounding these lesions is thought to precede the switch to an invasive phenotype. The molecular definition of angiogenesis is evolving, with a number of angiogenic growth and inhibitory factors adding to the complexity of the process. There is extensive evidence that vascular endothelial growth factor (VEGF) [10] plays a pivotal role in promoting angiogenesis and in tailoring a vascularized stroma. VEGF stimulates endothelial cell proliferation/migration and enhances vascular permeability [11,12]. The main isoforms of the protein (VEGF 121, VEGF 165, VEGF 189, and VEGF 206 ) bind to endothelial cells via at least two specific tyrosine kinase receptors, FLT-1 and KDR [13,14]. Moreover, the expression of both VEGF and its receptors is increased in several tumours [15 18]. Tumour angiogenesis and the role of angiogenic factors in human invasive breast cancer have been Copyright 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

2 Angiogenesis in pre-invasive breast lesions 141 extensively studied [19 23], while less information is available on their involvement in the early phases of breast neoplastic progression [7,24 27]. A group of epithelial pre-invasive lesions has been described, including florid ductal hyperplasia of usual type (FDHUT), atypical ductal hyperplasia (ADH), atypical lobular hyperplasia (ALH), welldifferentiated (WDIC), intermediately differentiated (IDIC) and poorly differentiated (PDIC) ductal carcinoma in situ, and lobular carcinoma in situ (LCIS). The nature of these lesions has yet to be elucidated, although some studies suggest that they are precursors of invasive breast cancer [28,29]. Ductal carcinoma in situ appears to be associated with substantial stromal angiogenesis and VEGF expression [25 27]. Similarly, some hyperplastic ductal lesions have shown a degree of stromal neovascularization [24,30,31]. However, lobular pre-invasive lesions have not been investigated exhaustively. The aim of the present study was to quantify angiogenesis and VEGF expression around all types of pre-invasive human breast lesion and to compare both parameters in normal glandular structures. (A) Materials and methods This study was approved by the Ethical Committee of the Pisa University Hospital. Specimens Samples of pre-invasive lesions (epithelial hyperplasia and carcinoma in situ) were obtained from breast biopsies or quadrantectomies of 136 women (age range years). Lesions were not associated with invasive carcinoma. Normal lobules were also analysed. They are termed collateral when present in the same sections hosting the pre-invasive lesions and non-collateral when obtained from normal mammary glands of nine young women (age range years), who underwent breast reduction [32]. Hyperplastic lesions were classified according to Page and Dupont [28] and Tavassoli and Norris [29]; ductal carcinomas in situ according to Holland et al [33]; and type A lobular carcinomas in situ (LCIS) according to Haagensen et al [34]. Type A LCIS was chosen on the basis of its strict morphological relationships with atypical lobular hyperplasia. In particular, in type A carcinoma, tumour cells have a monomorphic appearance and present a scant quantity of cytoplasm and small, round nuclei without nucleoli (Figure 1A). On the other hand, in ALH, epithelial cells are similar to those of type A LCIS, but the involvement of the lobule is less than 50% (Figure 1B). A total of 136 pre-invasive lesions were studied, including epithelial hyperplasia (40 FDHUT, 10 ADH, and 13 ALH) and ductal (20 PDIC, 25 IDIC, and 16 WDIC) and lobular (12 type A LCIS) carcinoma (B) Figure 1. (A) Type A lobular carcinoma in situ with acini completely distended by monomorphic cells with scant cytoplasm. (B) Atypical lobular hyperplasia showing a lobule with less than 50% involvement by similar cells in situ. A total of 672 normal structures (600 collateral and 72 non-collateral lobules) were also analysed. Tissues were fixed in 10% formalin and embedded in paraffin wax. Five-micrometre sections were obtained. Consecutive sections were used alternately for the haematoxylin and eosin (H&E) staining and for the immunohistochemistry to obtain a better comparison between morphology and protein expression. Immunohistochemistry (IHC) Sections were incubated with mouse anti-human CD34 (QB-END10; Novocastra Lab; dilution 1 : 100) and rabbit anti-human VEGF (Oncogene Research Products, Cambridge, MA, USA; dilution 1 : 100) primary antibodies, and detected as previously described [35]. Evaluation of parameters Vascularization Adjacent sections were either stained with H&E for histological evaluation or analysed by IHC to highlight the vessels. Following the criteria of Teo et al [7], a microvessel was defined as a single cell or a cluster of cells positive for CD34, sitting around a visible lumen

3 142 PViacavaet al clearly separate from adjacent microvessels and from other connective tissue components. In both lesions and normal structures, we were able to identify two distinct vascular patterns, previously described by Guidi et al and defined as pattern I and pattern II [25,27]. Pattern I, quantifiable as MVD (1) Ductal and lobular lesions: this indicates the diffuse presence of microvessels in the stroma surrounding the structures, in an area 500 µm wide. (2) Normal lobules: this indicates the diffuse presence of microvessels in the stroma of the structures. An average of five collateral normal lobules were analysed for each pathological sample. To avoid overlapping of vascular fields, only lobules at least 1 mm distant from pre-invasive lesions were selected. The analysis was performed at 200 ( 20 objective lens and 10 ocular lens) magnification, as described in a previous paper describing the microvascularization of normal human breast [32]. Microvessels were counted using an image analyser: images were digitized in a pixel matrix, using a colour video camera TK-1280E (JVC, Tokio, Japan) and a microcomputer processor. Digitized pictures were visualized on a high-resolution colour display (SAMPO, Tao-Yuan-Hsien, Taiwan). The true colour image analysis software package KS300 v.1.2 (Kontron Elektronik GmbH, Eching, Germany) was run for interactive manipulation, quantification, and data collection. The areas of lobular components were calculated by computer analysis and expressed as µm 2. Selected vessels were marked and numbered. MVD, expressed as vessel number/mm 2, was calculated as follows: [microvessel number/(microvessel area + residual stromal area)] independently and inconsistencies were resolved by a simultaneous review. Statistical analysis Statistical analysis was performed by the Mann Whitney test, using Statgraphics 5.0 (Statistical Graphics Corporation). Differences in the vascular parameters were considered statistically significant when p values were less than Results Vascularization Pattern I Ductal lesions Mean values of MVD around preinvasive ductal lesions and collateral normal parenchyma are reported in Figure 2. The highest values were observed in IDIC (mean value ± SD: 167 ± 61) and PDIC (190 ± 65), whereas in FDHUT and ADH MVD was significantly lower (p < 0.05) (123 ± 40 and 122 ± 47, respectively). In WDIC too, MVD was lower than in IDIC and PDIC (127 ± 46), although the difference was statistically significant only for PDIC (p < 0.001). No significant difference was found amongst FDHUT, ADH, and WDIC. Figure 3 shows an example of the appearance of pattern I in an IDIC sample. Lobular lesions MVD values in ALH were significantly lower than those in LCIS (125 ± 40 and 168 ± 50, respectively; p < 0.05) (Figure 2). Pattern II: not quantifiable, but given as absent or present This indicates the presence of a prominent cuff of microvessels in immediate apposition to the basement membrane of a single duct or acinus, both normal and pathological. It is considered present when either one structure is completely surrounded by a subtle rim of microvessels, or two or more structures are discontinuously surrounded by at least 50% of their circumference. VEGF expression VEGF expression was evaluated semi-quantitatively. The positivity index was obtained by counting all epithelial cells in normal breast structures or in preinvasive lesions, and by calculating the percentage of cells with VEGF cytoplasmic immunoreactivity, regardless of staining intensity. VEGF expression was determined by two pathologists (PV and AGN) Figure 2. Microvessel density (MVD; pattern I of vascularization) in pre-invasive breast lesions and collateral normal parenchyma. Dark grey bars: pre-invasive lesions; light grey bars: normal parenchyma collateral to pre-invasive lesions. FDHUT = florid ductal hyperplasia usual type; ADH = atypical ductal hyperplasia; WDIC = well-differentiated intraductal carcinoma; IDIC = intermediately differentiated intraductal carcinoma; PDIC = poorly differentiated intraductal carcinoma; ALH = atypical lobular hyperplasia; LCIS = lobular carcinoma in situ

4 Angiogenesis in pre-invasive breast lesions 143 Figure 3. Pattern I of vascularization in an intermediately differentiated intraductal carcinoma (IHC, anti-cd34) Normal parenchyma Lobules collateral to preinvasive lesions showed a delicate vascular framework. Overall, MVD of collateral lobules was lower than in pre-invasive lesions (115 ± 8 and 146 ± 26, respectively). However, single differences proved to be statistically significant only for IDIC, PDIC, and LCIS (p < 0.05). No differences were found when normal collateral and non-collateral structures were compared (115 ± 8 and 122 ± 18, respectively). Pattern II Lesions Pattern II was observed in all types of lesions, with variable percentages. A significant difference (p < 0.001) was found only when ductal lesions (22 ± 3.4% of positive samples) were compared with lobular lesions (3 ± 0% of positive samples) (Table 1). Figure 4 shows an example of the appearance of pattern II in a PDIC sample. Normal parenchyma A discontinuous rim of microvessels was present around the basement membrane of the acini within normal lobules. Table 1. Pattern II of vascularization in pre-invasive breast lesions Lesions % of cases positive Ductal FDHUT 20% ADH 21% WDIC 20% IDIC 28% PDIC 21% Lobular ALH 3% LCIS 3% FDHUT = florid ductal hyperplasia usual type; ADH = atypical ductal hyperplasia; WDIC = well-differentiated intraductal carcinoma; IDIC = intermediately differentiated intraductal carcinoma; PDIC = poorly differentiated intraductal carcinoma; ALH = atypical lobular hyperplasia; LCIS = lobular carcinoma in situ. Figure 4. Pattern II of vascularization in a poorly differentiated intraductal carcinoma. (IHC, anti-cd34) Patterns I and II No significant correlation was found when comparing the profiles of the two patterns. VEGF expression Normal parenchyma Luminal epithelial cells of lobules were characterized by heterogeneous VEGF expression. The range of positive cells varied between 2% and 4% (mean value 2.5 ± 0.43%), regardless of the type of collateral lesion. Lesions All types of lesion were characterized by heterogeneous VEGF expression. In ductal lesions, the number of VEGF-positive cells (13.5 ± 2%) was significantly (p < 0.001) higher than in lobular lesions (7.2 ± 2%) (Table 2). No significant difference was observed amongst the various types of ductal lesion, not even between ALH and LCIS. Exhaustive analysis of lesions showed no correlation between VEGF expression and vascularization pattern I or II. Overall, VEGF expression in lesions was significantly higher Table 2. VEGF expression in pre-invasive breast lesions Structures % of positive cells (mean ± SD) Ductal FDHUT 14.4 ± 9.2 ADH 10.2 ± 4.6 WDIC 14 ± 7.5 IDIC 15.3 ± 9.4 PDIC 14 ± 8.2 Lobular ALH 5.7 ± 5.5 LCIS 8.7 ± 7.2 FDHUT = florid ductal hyperplasia usual type; ADH = atypical ductal hyperplasia; WDIC = well-differentiated intraductal carcinoma; IDIC = intermediately differentiated intraductal carcinoma; PDIC = poorly differentiated intraductal carcinoma; ALH = atypical lobular hyperplasia; LCIS = lobular carcinoma in situ.

5 144 PViacavaet al Figure 5. Neoplastic cells immunoreactive for VEGF (arrows) in a poorly differentiated intraductal carcinoma (IHC, anti-vegf) than in normal lobules (10.4 ± 6.6% and 2.5 ± 0.4%, respectively; p < 0.001). Figure 5 shows an example of positivity for VEGF in a PDIC sample. Discussion The present paper analyses the angiogenic process within the full spectrum of pre-invasive lesions, both ductal and lobular, of the human breast. We found vascularization of the normal mammary gland and of lesions to be characterized by two distinctive morphological patterns, called pattern I and pattern II, as previously described [25]. The results are discussed separately for each pattern. Pattern I Angiogenesis and normal breast The degree of vascularization in normal glandular structures was found to be lower than in lesions. MVD values were consistent with those previously reported by our group in a set of samples from normal human mammary glands [32]. Remarkably, no difference was found when MVDs in normal collateral and noncollateral structures were compared. Since it cannot be excluded that the initiation phase of transformation had already taken place in some collateral lobules, we are inclined to rule out the hypothesis that neoangiogenesis is required during cancer initiation. We would rather believe, as discussed later, that it is required when promotion takes place, increasing gradually with tumour progression. Angiogenesis and epithelial hyperplasia (usual type and atypical) A first interesting result is the presence of neoangiogenesis in FDHUT, notwithstanding the low epidemiological and morphological risks associated with this lesion. Ottinetti and Sapino [30] reported increased mean vascular size in ductal hyperplasia, but no increase in MVD. Fregene et al [24] showed no difference in vascularity between proliferative and non-proliferative benign lesions, but found a significant increase in vascularity in proliferative lesions from post-menopausal women. Guinebretière et al [31] showed that the risk for hyperplasia to develop into breast cancer correlates with MVD. In FDHUT, ADH, and ALH, MVD values are intermediate between those of normal glandular structures and those of neoplastic lesions. These data indicate that an increase in blood supply is necessary for the appearance of any type of epithelial proliferation. The fact that both the typical and the atypical forms of hyperplasia have the same mean value of vascular density gives weight to the hypothesis that hyperplastic lesions characterized by higher MVD values would bear a higher risk of progression. Recent data by Jones et al [36] indicate that a proportion of FDHUT are indeed due to clonal neoplastic proliferation. Angiogenesis and ductal carcinoma in situ Angiogenesis was more pronounced in moderately and poorly differentiated cancer cell populations. In fact, IDIC and PDIC showed the highest MVD values, drawing significantly away from those observed in FDHUT and ADH. Heffelfinger et al [37], Engels et al [27], and Lee et al [38] reported intense vascularization in IDIC and PDIC. Guidi et al [25] described strong neovascularization in high-grade ductal carcinoma in situ, whereas Ottinetti and Sapino [30] and Porter et al [39] reported an increase in mean microvessel size but no change in MVD. WDIC showed peculiar features, with a degree of vascularization closer to hyperplasia than to other histological types of in situ ductal cancer. This observation is in line with a previous hypothesis from our and other groups that IDIC and PDIC may originate de novo, without a mandatory transition through WDIC [40,41]. Angiogenesis and lobular carcinoma in situ Information on angiogenesis in lobular lesions is still weak. The only paper, to our knowledge, published on the subject reports lower MVD values in lobular carcinoma in situ than in ductal carcinoma in situ, although data were derived from a very limited number of cases [37]. Interestingly, we found that LCIS showed a high level of angiogenesis, with MVD values similar to IDIC and PDIC. Our findings confirm the peculiar behaviour of LCIS. While sharing features of good differentiation with low-grade ductal lesions [40] (low proliferative activity, low p53 expression, high bcl-2 expression), LCIS also retains features of poor prognosis typical of high-grade ductal lesions, such as high laminin receptor expression [42]. Angiogenesis and VEGF VEGF was confirmed to play a crucial role, although angiogenesis is likely to be under multifactorial regulation. VEGF was expressed in all kinds of epithelial

6 Angiogenesis in pre-invasive breast lesions 145 cells throughout the different lesions. However, no correlation was found with the level or type of vascularization. Normal glandular structures showed a low number of VEGF-positive cells, in agreement with previously reported data [22,26,43]. Ductal lesions showed the highest production of VEGF, in terms of positive cells, with no significant difference amongst histological types. At variance with Heffelfinger et al [43], we were unable to find a progressive increase in VEGF from ductal hyperplasia to carcinoma in situ. Furthermore, the number of VEGF-positive cells in ductal carcinoma in situ was lower when compared with data from Lee et al [22]. The number of VEGFpositive cells was significantly lower in lobular than in ductal lesions. Lee et al [22] obtained similar results in invasive lobular carcinoma, with values lower than in invasive and in situ ductal carcinoma. Pattern II Pattern II was present in all types of lesions, but was linked predominantly to ductal lesions, with a positivity of 20 28% as opposed to 3% positivity in lobular lesions. Similarly, it was observed around normal ducts (data not shown) with a higher frequency than in normal lobules. The existence of two distinct vascular patterns in pre-invasive lesions suggests the possibility of two different angiogenic pathways [22]. Pattern II appears to be under the control of angiogenic factors released by neoplastic cells, while pattern I would be predominantly regulated by angiogenic factors from stromal inflammatory cells, particularly macrophages [44]. The fact that we are not able to demonstrate a correlation between VEGF production and the degree/type of vascularization supports the hypothesis that regulation of angiogenesis is multifactorial. A number of factors, produced by neoplastic and stromal inflammatory cells, have recently been described in pre-invasive breast lesions. Stromal vascularity in ductal carcinoma in situ was associated with thymidine phosphorylase, an angiogenic factor produced by perivascular inflammatory and stromal cells [45,46]. In addition, other angiogenic factors produced mainly by stromal cells, such as bfgf, PDGF-B, and IGFII, appear to cooperate with VEGF in vessel formation around preinvasive lesions [43]. We were unable to confirm a predominant role for stromal cells since we found very few stromal cells collateral to lesions positive for VEGF (data not shown), in agreement with previous reports [22]. The association of a new and rich vascularization with epithelial hyperplasia and in situ carcinoma has clinical implications. It provides the biological framework for using imaging techniques, such as MRI, to allow early detection of breast lesions on the basis of their increased vascularization. Furthermore, the early appearance of vessels could be exploited for prophylactic strategies in high-risk patients, such as women carrying BRCA1/2 gene mutations. The well-developed vascular network around pre-invasive lesions would either represent a target for antiangiogenic therapies or could be used as a vehicle for anti-proliferative drugs. Acknowledgements This study was supported by CNR-MIUR target project Molecular Diagnosis in Oncology and by AIRC, Italian Association for Cancer Research. References 1. Horak ER, Russel L, Klenk N, et al. Angiogenesis, assessed by platelet/endothelial-cells adhesion molecule antibodies, as indicator of node metastases and survival in breast cancer. Lancet 1992; 340: Giatromanolaki A, Koukourakis M, O Byrne K, et al. Prognostic value of angiogenesis in operable non-small-cell lung cancer. J Pathol 1996; 179: Toi M, Kashitani J, Tominaga T. Tumor angiogenesis is an independent prognostic indicator in primary breast carcinoma. Int J Cancer 1993; 55: Fina L, Molgaard H, Robertson D, et al. Expression of the CD34 gene in vascular endothelial cells. Blood 1990; 75: Vermeulen PB, Gasparini G, Fox SB, et al. Quantification of angiogenesis in solid human tumours: an international consensus on the methodology and criteria of evaluation. Eur J Cancer 1996; 32: Fontanini G, Lucchi M, Vignati S, et al. Angiogenesis as a prognostic indicator of survival in non small cell lung carcinoma: a prospective study. J Natl Cancer Inst 1997; 89: Teo NB, Shoker BS, Jarvis C, Martin L, Sloane JP, Holcombe C. Angiogenesis and invasive recurrence in ductal carcinoma in situ of the breast. Eur J Cancer 2003; 39: Smith MK, Weidner N. Demonstration and characterization of the angiogenic properties of cervical dysplasia. Cancer Res 1994; 54: Chodak GW, Hospelhorn V, Gittes R, Folkman J. Angiogenic activity as a marker of neoplastic and pre-neoplastic lesions of the human bladder. Ann Surg 1981; 192: Ferrara N, Henzel WJ. Pituitary follicular cells secrete a novel heparin-binding growth factor specific for vascular endothelial cells. Biochem Biophys Res Commun 1989; 161: Ferrara N. The role of vascular endothelial growth factor in pathological angiogenesis. Breast Cancer Res Treat 1995; 36: Kolch W, Martiny-Baron G, Kieser A, Marm D. Regulation of the expression of the VEGF/VPS and its receptors: role in tumor angiogenesis. Breast Cancer Res Treat 1995; 36: Shibuya M, Yamaguchi S, Yamane A, et al. Nucleotide sequence and expression of a novel human receptor-type tyrosine kinase gene (flt) closely related to the fms family. Oncogene 1990; 5: Terman BI, Carrion ME, Kovacs E, Rasmussen BA, Eddy RL, Shows TB. Identification of a new endothelial cell growth factor receptor tyrosine kinase. Oncogene 1991; 6: Alvarez JA, Baird A, Tatum A, et al. Localization of basic fibroblast growth factor and vascular endothelial growth factor in human glial neoplasms. Mod Pathol 1992; 5: Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 1995; 146: Brown LF, Berse B, Jackman RW, et al. Increased expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in kidney and bladder carcinomas. Am J Pathol 1993; 143:

7 146 PViacavaet al 18. Brown LF, Berse B, Jackman RW, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in adenocarcinomas of the gastrointestinal tract. Cancer Res 1993; 53: Bosari S, Lee AK, DeLellis RA, Wiley BD, Heatly GJ, Silverman ML. Microvessel quantification and prognosis in invasive breast carcinoma. Hum Pathol 1992; 23: Axelsson K, Ljung BME, More DH II, et al. Tumor angiogenesis as a prognostic assay for invasive ductal breast carcinoma. J Natl Cancer Inst 1995; 87: Brown LF, Berse B, Jackman RW, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in breast cancer. Hum Pathol 1995; 26: Lee AHS, Dublin EA, Bobrow LG, Poulsom R. Invasive lobular and invasive ductal carcinoma of the breast show distinct patterns of vascular endothelial growth factor expression and angiogenesis. J Pathol 1998; 185: De Jong JS, van Diest PJ, van der Valk, Baak JPA. Expression of growth factors, growth inhibiting factors, and their receptors in invasive breast cancer. I: Inventory in search of autocrine and paracrine loops. J Pathol 1998; 184: Fregene TA, Kellogg CM, Pienta KJ. Microvessel quantification as a measure of angiogenic activity in benign breast tissue lesions: a marker for precancerous disease? Int J Oncol 1994; 4: Guidi AJ, Fischer L, Harris JR, Schnitt SJ. Microvessel density and distribution in ductal carcinoma in situ of the breast. J Natl Cancer Inst 1994; 86: Guidi AJ, Schnitt SJ, Fischer L, et al. Vascular permeability factor (vascular endothelial growth factor) expression and angiogenesis in patients with ductal carcinoma in situ of the breast. Cancer 1997; 80: Engels K, Fox SB, Whitehouse RM, Gatter KC, Harris AL. Distinct angiogenic patterns are associated with high-grade in situ ductal carcinomas of the breast. J Pathol 1997; 181: Page DL, Dupont WD. Anatomic markers of human premalignancy and risk of breast cancer. Cancer 1990; 66: Tavassoli FA, Norris HJ. A comparison of the results of longterm follow-up for atypical intraductal hyperplasia and intraductal hyperplasia of the breast. Cancer 1990; 65: Ottinetti A, Sapino A. Morphometric evaluation of microvessels surrounding hyperplastic and neoplastic mammary lesions. Breast Cancer Res Treat 1988; 11: Guinebretière JM, Lé Monique G, Gavoille A, Bahi J. Angiogenesis and risk of breast cancer in women with fibrocystic disease. J Natl Cancer Inst 1994; 86: Naccarato AG, Viacava P, Bocci G, et al. Definition of the microvascular pattern of the normal human adult mammary gland. J Anat 2003; 203: Holland R, Peterse JL, Millis RR, et al. Ductal carcinoma in situ: a proposal for a new classification. Semin Diagn Pathol 1994; 11: Haagensen CD, Lane N, Lattes R. Neoplastic proliferation of the epithelium of the mammary lobules: adenosis, lobular neoplasia and small cell carcinoma. Surg Clin North Am 1972; 52: Viacava P, Gasperi M, Acerbi G, et al. Microvascular density and vascular endothelial growth factor expression in normal pituitary tissue and pituitary adenomas. J Endocrinol Invest 2003; 26: Jones C, Merrett S, Thomas VA, Barker TH, Lakhani SR. Comparative genomic hybridization analysis of bilateral hyperplasia of usual type of the breast. J Pathol 2003; 199: Heffelfinger SC, Yassin R, Miller MA, Lower E. Vascularity of proliferative breast disease and carcinoma in situ correlates with histological features. Clin Cancer Res 1996; 2: Lee AH, Happerfield LC, Bobrow LG, Millis RR. Angiogenesis and inflammation in ductal carcinoma in situ of the breast. J Pathol 1997; 181: Porter PL, Patton KF, Self SA, et al. A quantitative study of blood vessels size and density in normal and neoplastic breast tissue. USCAP abstract #87. Mod Pathol 1993; 6: Viacava P, Naccarato AG, Bevilacqua G. Different cell growth patterns characterize different preinvasive breast lesions. J Pathol 1999; 188: Buerger H, Otterbach F, Simon R, et al. Comparative genomic hybridization of ductal carcinoma in situ of the breast evidence of multiple genetic pathways. J Pathol 1999; 187: Viacava P, Naccarato AG, Collecchi P, Mènard S, Castronovo G, Bevilacqua G. The spectrum of 67-kD laminin receptor expression in breast carcinoma progression. J Pathol 1997; 182: Heffelfinger SC, Miller MA, Yassin R, Gear R. Angiogenic growth factors in preinvasive breast disease. Clin Cancer Res 1999; 5: Lewis JS, Landers RJ, Underwood JCE, Harris AL, Lewis CE. Expression of vascular endothelial growth factor by macrophages is up-regulated in poorly vascularized areas of breast carcinoma. J Pathol 2000; 192: Engels K, Fox SB, Whitehouse RM, Gatter KC, Harris AL. Upregulation of thymidine phosphorylase expression is associated with a discrete pattern of angiogenesis in ductal carcinoma in situ of the breast. J Pathol 1997; 182: Lee AHS, Dublin EA, Bobrow LG. Angiogenesis and expression of thymidine phosphorylase by inflammatory and carcinoma cells in ductal carcinoma in situ of the breast. J Pathol 1999; 187: