- Email: [email protected]
Prognostic significance of microvessel density and vascular endothelial growth factor expression in sinonasal carcinomas
Human Pathology (2006) 37, 391 – 400
www.elsevier.com/locate/humpath
Original contributions
Prognostic significance of microvessel density and vascular endothelial growth factor expression in sinonasal carcinomasB Guido Valente MDa,*, Carlo Mamo MDc, Antonella Bena MDc, Elisa Prudente BTechna, Cristina Cavaliere BScd, Simonetta Kerim MDd, Giuseppina Nicotra BScb, Alberto Comino MDe, Giorgio Palestro MDd, Ciro Isidoro MDb, Fabio Beatrice MDf a
Pathology Section, Department of Medical Sciences, Amedeo Avogadro University Medical School, I 28100 Novara, Italy Laboratory of Molecular Pathology, Department of Medical Sciences, Amedeo Avogadro University Medical School, I 28100 Novara, Italy c Epidemiology Unit, Agenzia Regionale Protezione Ambientale, I 10095 Grugliasco, Italy d Department of Biomedical Sciences and Human Oncology, School of Medicine and San Giovanni Hospital, I 10100 Torino, Italy e Laboratory of Anatomic Pathology, Santa Croce and Carle Hospital, I 12100 Cuneo, Italy f Department of Otolaryngology, Giovanni Bosco Hospital, I 10100 Torino, Italy b
Received 9 August 2005; revised 18 November 2005; accepted 21 November 2005
Keywords: Sinonasal carcinoma; Microvessel density; Tumor angiogenesis
Summary The prognostic significance of microvessel density and proliferative activity of the neoplastic cells, evaluated respectively by CD31 and Ki-67 positivity, and immunohistochemical expression of vascular endothelial growth factor (VEGF) was retrospectively investigated in 105 cases of sinonasal carcinoma (80 surgical specimens and 25 biopsies). The most represented histologic types were intestinal-type adenocarcinoma found in 36 patients (34.3%), squamous cell carcinoma (SCC) in 34 (32.4%), mucinous adenocarcinoma (mainly made up of signet-ring cell patterns) in 15 (14.3%), and adenoid cystic carcinoma in 7 (6.7%). Microvessel density values (in vessels per square millimeter), VEGF, and Ki-67 were not dependent on histologic type but were rather correlated to the histologic grading in SCC. Clinical data were available for 92 (87.6%) of 105 patients, with minimum follow-up of 48 months. Most of the patients (81.5%) were at an advanced stage (T3-T4) at diagnosis. The values of all markers were correlated to tumor stage ( P = .03). Multivariate analysis showed that both microvessel density and proliferative activity of the neoplastic cells were independent prognostic parameters (mortality hazard ratio, 1.33 and 1.60, respectively). Although VEGF expression was not correlated to prognosis on the whole series ( P = .06), it was a powerful prognostic marker when the analysis was restricted to the group of SCCs (hazard ratio, 3.02; 90% confidence interval, 1.58-5.80). These results show that tumor neoangiogenesis, expressed by microvessel density, together with proliferative activity, is a pathologic marker with a strong prognostic impact in sinonasal carcinomas. Therefore, it may be a
B
This work was supported by a grant from the Piedmont Region. * Corresponding author. E-mail address: [email protected] (G. Valente). 0046-8177/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2005.11.021
392
G. Valente et al. useful tool in this field so as to carry out therapeutic protocol planning, which may be further enhanced by the adoption of the more recent antiangiogenic molecules. D 2006 Elsevier Inc. All rights reserved.
1. Introduction When evaluating tumor growth, neoangiogenesis is known to be one of the essential events in neoplastic progression. A great deal of interest has been paid to the predictive value of neoangiogenesis (represented as microvessel density determined by the evaluation of CD31, CD34, and FVIIIr positivity) on the clinical progression and prognosis of several types of tumors over the last decade [1]. Indeed, it has been demonstrated that microvessel density has a prognostic role for a number of epithelial tumors, including carcinomas of the breast [2], lung [3], and large bowel [4], as well as for melanoma [5] and gliomas [6]. The regulation of the neoangiogenetic process is complex; normal tissues have a number of factors with proangiogenic or antiangiogenic function, produced by stromal and inflammatory cells, which are strictly balanced in normal tissues, whereas proangiogenic factors are usually prevalent in neoplastic tissues. Immunohistochemical and/or molecular techniques have been used to demonstrate a relationship between neoangiogenesis and expression of growth factors with proangiogenetic activity, such as vascular endothelial growth factor (VEGF) and its subunits [7]. The expression of these molecules has been seen to have a prognostic significance in many epithelial and nonepithelial tumors. This study analyzes the microvessel density and the growth fraction (by CD31 and Ki-67 positivity, respectively), as well as the immunohistochemical expression of VEGF, in a multicentric series of 105 cases of carcinomas that developed in the nose and paranasal sinuses, with the aim of assessing their prognostic role. These tumors are relatively infrequent and often related to occupational exposures, and to date, their prognosis has been based on traditional clinical and pathologic features. TNM staging and histologic type were seen to be particularly significant, as it has been observed that patients with T3 to T4 stages and squamous cell or undifferentiated carcinoma have a worse outcome, as also recently documented [8], whereas glandular tumors seem to be associated with a better prognosis. Moreover, localization and treatment modality also show an influence on the actuarial survival rate. Because diagnosis is frequently made at an advanced stage and the outcome remains ominous in many cases, as an average of only 40% of these patients have a 5-year survival after first diagnosis [8], a need to acquire biologic markers such as those analyzed in the present study clearly exists to estimate the aggressiveness of these tumors with as much accuracy as possible. Furthermore, data on the neoangiogenesis may also have therapeutic implications
when the latest reports about the application of antiangiogenic substances in solid tumors are taken into consideration.
2. Materials and methods 2.1. Patients The study was conducted on 105 cases of malignant epithelial tumors of nasal cavity and paranasal sinuses, mainly obtained from the files of the pathology departments of the University of Torino (San Giovanni Hospital) and Novara and from Cuneo Hospital (Italy). The sources of the other cases have been listed in the acknowledgment section. Table 1 Main clinicopathologic data of the 105 cases included in this study Sex Age (y)
Male Female 20-29 30-39 40-49 50-59 60-69 70-79 N80
Histologic type Intestinal-type adenocarcinoma Grade 1 Grade 2 Grade 3 SCC Grade 1 Grade 2 Grade 3 Mucinous adenocarcinoma Adenoid cystic carcinoma Oncocytoid adenocarcinoma Low-grade NOS adenocarcinoma Basaloid SCC Schneiderian carcinoma Verrucous carcinoma Others (low-grade mucoepidermoid carcinoma, small cell carcinoma, teratocarcinosarcoma) Surgical therapy Yes No Abbreviation. NOS, not otherwise specified.
n
%
84 21 2 8 18 47 19 9 2
80.0 20.0 1.9 7.6 17.1 44.8 18.1 8.6 1.9
36 4 23 9 34 0 13 21 15 7 2 2
34.3
2 2 2 3
1.9 1.9 1.9 2.8
80 25
76.2 23.8
32.4
14.3 6.7 1.9 1.9
Microvessel density and VEGF expression in carcinomas Table 2 Clinicopathologic findings in 92 cases with available clinical files Staging T1 T2 T3 T4 Surgical therapy Yes No Chemotherapy Yes No Radiotherapy Yes No Metastasis Yes No Relapse Yes No
n
%
1 16 22 53
1.1 17.4 23.9 57.6
74 18
80.4 19.6
22 70
23.9 76.1
80 12
86.9 13.1
7 85
7.6 92.4
41 51
44.6 55.4
The main clinicopathologic data are summarized in Table 1. The study cohort included 84 men and 21 women, with an average age of 64.3 F 11.1 years (range, 28-83 years). Surgical approach, followed by postoperative radiotherapy and/or chemotherapy, was the first choice therapy in 80 cases (76.2%). The remaining 25 patients were not operable and were treated exclusively by radiochemotherapy. All routine histologic slides were reviewed and histologic diagnoses were reformulated, according to the World Health Organization [9]: intestinal-type adenocarcinoma (36 cases) was the most represented histologic variety, followed by 34 squamous cell carcinomas (SCCs), 15 mucinous adenocarcinomas (with prevalence of signet-ring cell patterns), and 7 adenoid cystic carcinomas. More unusual tumors were recognized in the other 13 cases. The separation of intestinal-type adenocarcinomas and mucinous adenocarcinomas was justified by the worse prognosis demonstrated by the latter. The histologic grading was also reformulated, according to the World Health Organization criteria, for SCC [9]. Because of its low reproducibility, histologic grading of adenocarcinoma was not formulated. Both clinical files and radiologic documentation were available to help in the new determination of the clinical staging at first diagnosis in 92 patients (72 men and 20 women) (Table 2): 1 patient was reclassified as stage I, 16 patients (17.4%) as stage II, 22 (23.9%) as stage III, and 53 (57.6%) as stage IV. Follow-up ranged from 48 to 96 months.
2.2. Immunohistochemical studies The following monoclonal antibodies were used on paraffin sections: clone JC70A (DAKO A/S, Glostrup,
393 Denmark) to identify CD31, clone MIB-1 (DAKO A/S) to identify Ki-67, and clone C-1 (Santa Cruz Biotechnology, Santa Cruz, Calif) and clone 26503.11 (Sigma, St Louis, Mo) to identify human VEGF. The anti-VEGF monoclonal antibodies were raised against the 140 and 165 amino acids isoforms, respectively. The sections had previously been placed in 10 mmol/L citrate buffer, pH 6, and microwaved at 750 W for two 5-minute periods. Immunostaining was performed using the labeled streptavidin-biotin method (LSAB2 prediluted system kit, DAKO A/S), and diaminobenzidine 3,3-tetrachloride was used as chromogen. Negative controls omitted the primary antibody. The evaluation of CD31 positivity was restricted to microvessels inside the intratumoral stroma or on the front of invasion, according to the bhot spotQ method [10]. At first, 10 fields for each case showing a high microvessel density were selected at low magnification; the vessels were then counted at high magnification (40 objective) in each field. The square of the microscopic field was determined, and the average value was compared with 1 mm2 extension for each case. Three groups of cases were arbitrarily individualized on the basis of the results: less than 100, 100 to 150, and more than 150 vessels per square millimeter. Vascular endothelial growth factor positivity was displayed in the cytoplasms of the neoplastic cells with a more or less strong intensity and was evaluated by the following semiquantitative method: , negative or only sporadic positive cells; +, 5% to 20% positive cells; ++, 20% to 70% positive cells; and +++, more than 70% positive cells (+ was identified with the number 1 (+), 2 (++), and 3 (+++) for the statistical evaluation). The nuclei of the neoplastic cells characteristically showed Ki-67 positivity. Ten high-magnification fields were evaluated for each case, and the percentage of the neoplastic positive cells was calculated. Finally, the average percentage of the positive cells was obtained. Three groups of cases
Fig. 1 Intestinal-type adenocarcinoma. CD31-positive neoformed vascular structures are particularly numerous on the front of invasion (immunoenzymatic method with streptavidin, original magnification 275).
394
G. Valente et al.
2.3. Specificity of the antihuman VEGF antibody The specificity of the anti-VEGF monoclonal antibody used in the present study was ascertained by using Western blotting of VEGF expression in cultured human colorectal cancer DLD1 cells. Cell homogenate was prepared by freeze-thawing cycles and sonication in a buffer containing 2% sodium deoxycholate and a cocktail of protease inhibitors. Fifty micrograms of protein was heat-denatured in a Laemmli buffer not containing b-mercaptoethanol, separated by sodium dodecyl sulfate–polyacrilamide gel electrophoresis, and then electroblotted onto nitrocellulose filters. Western blotting was performed after a standard procedure as described elsewhere [11] using the anti-VEGF as primary antibody and a peroxidase-conjugated rabbit antimouse IgG polyclonal antiserum as secondary antibody.
2.4. Statistical analysis Fig. 2 Squamous cell carcinoma. CD31-positive neoformed vessels are well represented close to neoplastic sheets (immunoenzymatic method with streptavidin, original magnification 275).
were arbitrarily individualized: less than 25% positive neoplastic cells, 25% to 50% positive cells, and more than 50% positive cells. All immunohistochemical reactions were evaluated by 2 persons (G. V. and S. K.) for each case; the interobserver difference was less than 10%. In a limited number of cases, sections were also processed for immunofluorescence staining. In this case, deparaffinized biopsies were incubated overnight at 48C with the primary antibody (diluted 1:50 in phosphate buffer containing 0.1% Triton X-100 and 4% fetal calf serum). After having washed out the excess of the antibody, sections were subsequently incubated for 1 hour at room temperature with a fluorescein isothiocyanate–conjugated goat antimouse polyclonal antiserum as the secondary antibody (diluted 1:350, as aforementioned). All incubations were performed in a humidified chamber protected from light. Negative control was performed by omitting the primary antibody. Sections were counterstained with propidium iodide (1:500 with the secondary antibody) to label the nucleus. Table 3
The v 2 statistic method was used to check the association between 2 variables (for instance microvessel density and tumor stage). Survival curves were estimated by the product-limit (Kaplan-Meier) method. The log-rank test was performed to compare 2 or more survival curves for different groups of individuals. Cox proportional hazards models were used to investigate associations of microvessel density, VEGF expression, and growth factor with diseasefree survival. Analysis was carried out by SAS statistical software (SAS Institute Inc, Cary, NC).
3. Results 3.1. CD31 immunostaining and microvessel density CD31 was displayed by endothelial cells as well as by a fraction of the tumor-infiltrating mononuclear cells. Vascularization was mainly localized in the stroma surrounding the neoplastic tissue and was prominent in the invasive front (Figs. 1 and 2). Microvessel counts varied from 50 to 232.95/mm2, with a median value of 121.8/mm2 (average, 123.9/mm2). The most represented histologic types had values ranging from 50.6 to 209.8 in intestinal-type adenocarcinomas (median value, 122.72), whereas they were
Positivity for CD31, VEGF, and growth fraction in the most frequent histologic types of sinonasal carcinoma Microvessel density (CD31) (/mm2)
Intestinal-type adenocarcinoma SCC Mucinous adenocarcinoma Adenoid cystic carcinoma
VEGF
Growth fraction (% positive cells)
b100
100-150
1 (+)
2 (++)
3 (+++)
b25
25-50
15 (41.7)
13 (36.1)
8 (22.2)
13 (36.2)
7 (19.4)
16 (44.4)
12 (33.3)
17 (47.2)
7 (19.5)
11 (32.4) 6 (40.0)
13 (38.2) 7 (46.7)
10 (29.4) 2 (13.3)
13 (38.2) 7 (46.7)
7 (20.6) 3 (20.0)
14 (41.2) 5 (33.3)
2 (5.9) 6 (40.0)
15 (44.1) 5 (33.3)
17 (50.0) 4 (26.7)
1 (14.3)
4 (57.1)
2 (28.6)
1 (14.3)
4 (57.1)
2 (28.6)
5 (71.4)
2 (28.6)
Values are presented as n (%).
N150
N50
0 (0)
Microvessel density and VEGF expression in carcinomas Table 4
Positivity for CD31, VEGF, growth fraction, and clinical stage in 31 cases of intestinal-type adenocarcinoma Microvessel density (CD31) (/mm2)
T1 T2 T3 T4
(0) (6) (11) (14)
395
VEGF
Growth fraction (% positive cells)
b100
100-150
N150
1 (+)
2 (++)
3 (+++)
b25
25-50
N50
/ 5 (83.3) 3 (27.3) 4 (28.6)
/ 1 (16.7) 5 (45.4) 6 (42.8)
/ / 3 (27.3) 4 (28.6)
/ 2 (33.4) 4 (36.4) 5 (35.7)
/ 2 (33.3) 2 (18.2) 2 (14.3)
/ 2 (33.3) 5 (45.4) 7 (50.0)
/ / 5 (45.4) 4 (28.6)
/ 3 (50.0) 3 (27.3) 7 (50.0)
/ 3 (50.0) 3 (27.3) 3 (21.4)
Values are presented as n (%). Slash (/) represents a zero value.
from 59.1 to 172.72 (median value, 106.81) in mucinous and signet-ring cell adenocarcinomas, from 50.0 to 224.24 (median value, 119.86) in SCC, and from 58.63 to 187.87 (median value, 125.0) in adenoid cystic carcinomas (Table 3). Although microvessel density did not show any significant correlation with the histologic type of the tumors, the v 2 test demonstrated a significant association between microvessel density values and the clinical stage ( P = .001) (Tables 4 and 5) and with poor histologic grading in SCC ( P = .04).
3.2. Vascular endothelial growth factor immunostaining The presence and localization of VEGF in biopsy sections of tumor specimen were analyzed using immunocytochemical methods. In Fig. 3, an example of immunofluorescence staining is shown; strong positive granules are localized in the cytoplasms of some neoplastic cells in a SCC. Extensive analysis of VEGF expression in all tumor specimens was performed by immunohistochemical staining using the streptavidin-biotin peroxidase method as previously described. No substantial differences in localization and intensity of the reaction product were observed between the 2 antibodies used, although clone C-1 showed a slightly higher background staining. Vascular endothelial growth factor positivity was almost completely restricted to the neoplastic tissue and was expressed in the cytoplasm of neoplastic cells in all cases (Figs. 4 and 5). There was a rather uniform distribution of the positive neoplastic cells in the tumor tissue. The semiquantitative evaluation showed a high number of positive cells (N75%) in 16 (44.4%) of 36 cases of intestinal-type adenocarcinomas, in 14 (41.1%) of 34 cases of SCC, and in 5 (33.3%) of 15 mucinous and signet-ring cell adenocarcinomas (Table 3). More than 50% of the neoplastic cells expressed VEGF in 6 of 7 adenoid cystic carcinoma (85.7%). The expression of VEGF was associated with microvessel Table 5
(0) (6) (7) (17)
3.3. Specificity of the anti-VEGF Specificity of the anti-VEGF antibody used in the present work was ascertained by Western blotting analysis of VEGF expression in cell homogenate of human colorectal cancer DLD1 cells. As shown in Fig. 6, the antibody revealed the presence of bands corresponding to the polypeptides of expected size for VEGF isoforms [12].
3.4. Growth fraction The growth fraction was evidenced by the nuclear positivity with Ki-67 antibody (Figs. 7 and 8). High levels of proliferative activity were observed in SCC, as more than 50% of Ki-67–positive neoplastic cells were counted in 17 (50%) of 34 cases, whereas similar values were observed in only 7 (19.4%) of 36 intestinal-type adenocarcinomas and 4 (26.6%) of 15 mucinous and signet-ring cell adenocarcinomas (Table 3). The growth fraction was also seen to be associated with clinical stage ( P = .03) (Tables 4 and 5) and histologic grading in SCC ( P b .0001).
3.5. Patients outcome Patients with high microvessel density (N150 vessels per square millimeter) showed a worse 5-year disease-free survival than those with intermediate (100-150/mm2) or low (b100/mm2) microvessel density (Fig. 9). These survival curves differed significantly even when adjusted for the effect of the histologic type (log-rank test, P = .03). Multivariate analysis revealed that microvessel density was a significant independent prognostic factor; mortality hazard ratio adjusted for sex, age, therapy, staging, and histologic
Positivity for CD31, VEGF, growth fraction, and clinical stage in 30 cases of SCC Microvessel density (CD31) (/mm2)
T1 T2 T3 T4
density ( P = .01) and with poor histologic grading in SCC ( P = .04). It was, moreover, associated with clinical stage ( P = .03) (Tables 4 and 5).
VEGF
Growth fraction (% positive cells)
b100
100-150
N150
1 (+)
2 (++)
3 (+++)
b25
25-50
N50
/ 5 (83.3) 1 (14.4) 3 (17.7)
/ 1 (16.7) 3 (42.8) 9 (52.9)
/ / 3 (42.8) 5 (29.4)
/ 2 (33.3) 3 (42.8) 7 (41.2)
/ 1 (16.7) / 5 (29.4)
/ 3 (50.0) 4 (57.2) 5 (29.4)
/ / 1 (14.4) 1 (5.9)
/ 3 (50.0) 2 (28.4) 9 (52.9)
/ 3 (50.0) 4 (57.2) 7 (41.2)
Values are presented as n (%). Slash (/) represents a zero value.
396
G. Valente et al.
Fig. 5 Squamous cell carcinoma. Neoplastic cells displaying intense VEGF positivity (immunoenzymatic method with streptavidin, original magnification 275). Fig. 3 Squamous cell carcinoma. Immunofluorescence staining with monoclonal antibody against human VEGF revealed with fluorescein isothiocyanate–conjugated secondary antibody (green staining). Nuclei are stained with propidium iodide (red staining). Vascular endothelial growth factor positivity is displayed by the cytoplasm of neoplastic cells (original magnification 200).
staging, and histologic type was 1.60 (90% CI, 1.06-2.42) for increasing level of Ki-67. Vascular endothelial growth factor positivity, when expressed semiquantitatively, was not related to the survival on the overall series. However, this marker was seen to have a high impact on the disease-free survival for the patients with SCC (log-rank test, P = .07) (Fig. 11).
type was 1.33 (90% confidence interval [CI], 0.90-1.95) for increasing level of microvessel density. When Ki-67 was categorized on 3 levels (b25%, 25%-50%, and N50% positive cells), patients with a high growth fraction (N50% of positive neoplastic cells) also showed a worse 5-year disease-free survival (log-rank test, P = .15) (Fig. 10). However, the prognosis was worse in the first group (b25%) than in the second (25%-50%) at 5-year follow-up. Ki-67 positivity was a significant independent prognostic factor; mortality hazard ratio adjusted for sex, age, therapy,
Fig. 4 Mucinous adenocarcinoma. Strong VEGF positivity in the cytoplasms of the neoplastic cells (immunoenzymatic method with streptavidin, original magnification 275).
Fig. 6 Specificity of antihuman VEGF antibody used in the present study. Vascular endothelial growth factor peptides were identified by Western blotting analysis. Molecular weights correspond to the expected size for dimeric forms of VEGF peptides made up of 121, 165, 189, and 206 amino acids. The 2 anti-VEGF monoclonal antibodies gave similar results.
Microvessel density and VEGF expression in carcinomas
397
4. Discussion
in agreement with previous observations on other tumors, that is, microvessel density, as measured by CD31 immunostaining, is a highly predictive and independent marker of both progression and prognosis. Neoangiogenesis is correlated to other progression parameters, mainly the clinical stage. Furthermore, the entity of the neoangiogenesis is not related to the intrinsic aggressiveness of the different histologic types because low microvessel densities were also evaluated in varieties known to have a worse prognosis, such as the signet-ring cell adenocarcinoma [23]. This may explain the powerful impact of microvessel density in our histologically heterogeneous series. There was a correlation between microvessel density and the immunohistochemical expression of VEGF (the most investigated of proangiogenic proteins); a similar result was observed in many other experimental and human tumors, for instance, in head and neck carcinomas [22,24], because the neovascularization should be directly up-regulated by VEGF and other proangiogenic factors. Moreover, in normal conditions, also VEGF seems to be implicated in the development of blood [25] and lymphatic vessels through its receptor, VEGFR3 (flt-4) [26]. Indeed, this is an interesting observation because lymphatic diffusion is prominent in head and neck tumors. Vascular endothelial growth factor proved to be a progression marker for SCC in our series, as demonstrated in other head and neck sites [27]. High amounts of VEGF have been found in supernatants of head and neck SCC [28], leading to the hypothesis of a more consistent role in these, rather than in other histologic types. However, apart from the predictive value found in the SCC group, the expression of VEGF lacked a significant prognostic value in our whole series as in other reports [29]. The discrepancy between this observation and the strong prognostic role of microvessel density, in spite of their relationship, may be explained by the fact that the expression of single
When evaluated by the hot spot method, microvessel density proved to be a valuable prognostic indicator in several human tumors. The validation of neovascularization as a relevant prognostic tool has been widely confirmed in most retrospective studies on breast carcinoma, non–small cell lung carcinoma, and prostate carcinoma, which showed that microvessel density is directly correlated with the clinical outcome in these tumors [13]. Positive results have also been obtained in several studies on gastric [14] and bladder carcinoma [15] and testicular germinal cell tumors [16], as well as hematologic malignancies [17]. Contradictory results on prognosis have been found in colorectal carcinomas [18,19] and in malignant melanoma of the skin [20,21], whereas in some tumors such as papillary thyroid carcinoma, tumor angiogenesis was correlated to the development or maintenance of the tumor but not to the outcome [22]. There is a shortage of data on the predictive value of microvessel density in sinonasal carcinomas. Our results are
Fig. 8 Squamous cell carcinoma. Neoplastic sheets are made up by cells with marked Ki-67 positivity (immunoenzymatic method with streptavidin, original magnification 275).
Fig. 7 Intestinal-type adenocarcinoma with high growth fraction. Most of the neoplastic cells show nuclear Ki-67 positivity (immunoenzymatic method with streptavidin, original magnification 275).
Multivariate analysis revealed that VEGF had a weak prognostic impact on the whole series; hazard ratio adjusted for sex, age, surgical therapy, staging, and histologic type was 1.10 (90% CI, 0.79-1.53) for increasing level of VEGF. Noteworthy was the fact that VEGF was seen to be an important prognostic factor in SCC; hazard ratio was 3.02 (90% CI, 1.58-5.80) in these patients because of the increase in VEGF level when adjusted for sex, age, therapy, staging, and histologic grading.
398
G. Valente et al.
Fig. 9 Kaplan-Meier survival curves by level of microvessel density (3 levels) in 92 cases of malignant epithelial tumors of nasal cavity show that survival is significantly worse in patients with microvessel density higher than 150/mm2. The differences between the curves has been evaluated by log-rank test ( P = .03).
proangiogenic or antiangiogenic factors is a dynamic, more than a static, biologic process, balanced by both positive and negative angiogenesis regulators [30], and may vary over time [31]. Analysis of the growth fraction, as determined by MIB-1 immunohistochemical staining, had a significant and independent prognostic role in sinonasal carcinomas. To the best of our knowledge at the time of writing, no relevant results have been reported as to what impact proliferative
activity has on these tumors. A direct relationship between proliferative activity and neoangiogenesis was expected because a good vascular support is necessary to warrant the growth of the neoplastic cells, particularly on the front of invasion [32]. The growth fraction represents a further strong prognostic evaluation element from a practical point of view. Correlations between biologic parameters and clinical outcome have rarely been reported in nasal tumors, and the
Fig. 10 Kaplan-Meier survival curves by level of growth fraction (3 levels of Ki-67 positivity). The 5-year disease-free survival is significantly worse in patients with more than 50% of positive neoplastic cells. After 3 years, patients with low growth fraction (b25%) show a worse prognosis than those with intermediate value (25%-50%). The differences between the curves has been evaluated by log-rank test ( P = .15).
Microvessel density and VEGF expression in carcinomas
399
Fig. 11 Survival curves restricted to cases of SCC demonstrate that VEGF expression is a powerful marker of prognosis in these tumors. Cases belonging to group 3 (N75% positive cells) show a negative 5-year disease-free survival. The differences between the curves has been evaluated by log-rank test ( P = .07).
prognosis of these tumors, also in the most recently reported series, has mainly been formulated on the basis of traditional parameters [8,33]. As expected, these later proved to be prognostically significant also in our series. However, the clinical application of our results could be of great relevance, in as much as the diagnosis of many tumors of nose and paranasal sinuses is often formulated too late to allow a surgical approach and/or alternative therapies such as exclusive radiotherapy that are able to reduce the neoplastic mass. Microvessel density has been demonstrated to be a useful tool in the prediction of radiosensitivity in laryngeal SCC [34], and antiangiogenic therapies used to reduce a neoplastic mass that cannot be operated on and are now applied in several clinical trials. Whether microvessel density may be an indicator of the susceptibility to the treatment with antiangiogenic molecules or alternatively a predictive marker of response is still debatable. Even if microvessel density is used to reflect the angiogenesis, the two parameters cannot be considered to completely overlap [35]. Moreover, although tumors with high microvessel density should be, from a rational point of view, the most suitable candidates for antiangiogenic therapy [36], there is experimental evidence that, in mice, less vascularized bladder tumors give the best response [37]. In conclusion, sinonasal carcinomas can be added to the many other malignant tumors where prognosis is related to the neoangiogenesis. Both microvessel density and growth fraction have emerged as being powerful prognostic and predictive tools in our study. These results could well indicate a further application in the light of therapeutic approach using antiangiogenic molecules that are potentially useful in these neoplasias that are often diagnosed at
an advanced stage for which an extrasurgical approach is preferred.
Acknowledgments The authors thank the following pathologists who contributed to the article by sending their cases: Dr Enrico Bollito (San Luigi Hospital, Orbassano, Italy), Dr Bruno Torchio and Dr Manuela Motta (Mauriziano Hospital, Torino, Italy), Dr Stefano Taraglio and Dr Laura Davico Bonino (Maria Vittoria Hospital, Torino, Italy), Dr Mauro Giudici (Ospedale degli Infermi, Biella, Italy), Dr Clemente Vineis (Ospedale Civile di Ivrea, Italy), Dr Mario Abrate (ASL 17 of Savigliano, Fossano e Saluzzo, Italy), and Dr Giovanni Cera (ASL 16 di Mondovı`, Italy). The authors thank Mrs Barbara Wade for revising the English language and Dr. Gabriella Nicosia and Mr Denis Longhi for the skillful technical assistance.
References [1] Hasan J, Byers R, Jayson GC. Intra-tumoural microvessel density in human solid tumours. Br J Cancer 2002;86:1566 - 77. [2] Gasparini G, Toi M, Verderio P, et al. Prognostic significance of p53, angiogenesis, and other conventional features in operable breast cancer: subanalysis in node-positive and node-negative patients. Int J Oncol 1998;12:1117 - 25. [3] Giatromanolaki A, Koukourakis MI, Kakolyris S, et al. Vascular endothelial growth factor, wild-type p53, and angiogenesis in early operable non–small cell lung cancer. Clin Cancer Res 1998;4:3017 - 24. [4] Banner BF, Whitehouse R, Baker SP, et al. Tumor angiogenesis in stage II colorectal carcinoma: association with survival. Am J Clin Pathol 1998;109:733 - 7.
400 [5] Neitzel LT, Neitzel CD, Magee KL, et al. Angiogenesis correlates with metastasis in melanoma. Ann Surg Oncol 1999;6:70 - 4. [6] Schmidt NO, Westphal M, Hagel C, et al. Levels of vascular endothelial growth factor, hepatocyte growth factor/scatter factor and basic fibroblast growth factor in human gliomas and their relation to angiogenesis. Int J Cancer 1999;84:10 - 8. [7] Hlatky L, Hahnfeldt P, Folkman J. Clinical application of antiangiogenic therapy: microvessel density, what it does and doesn’t tell us. J Natl Cancer Inst 2002;94:883 - 93. [8] Dulguerov P, Jacobsen MS, Allal AS, et al. Nasal and paranasal sinus carcinoma: are we making progress? A series of 220 patients and a systematic review. Cancer 2001;92:3012 - 29. [9] Shanmugaratnam K, Sobin LH, editors. Histological typing of tumours of the upper respiratory tract and ear. 2nd ed. WHO international histological classification of tumours. Berlin7 SpringerVerlag; 1991. p. 32. [10] Weidner N, Semple JP, Welch WR, et al. Tumor angiogenesis and metastasis—correlation in invasive breast carcinoma. N Engl J Med 1991;324:1 - 8. [11] Isidoro C, De Stefanis D, Demoz M, et al. Expression and posttranslational fate of Cathepsin D in HT-29 tumor cells depend on their enterocytic differentiation state. Cell Growth Differ 1997;8:1029 - 37. [12] Ferrara N, Henzel WJ. Pituitary follicular cells secret a novel heparinbinding growth factor specific for vascular endothelial cells. Biochem Biophys Res Commun 1989;161:851 - 8. [13] Gasparini G, Harris AL. Prognostic significance of tumor vascularity. In: Teicher BA, editor. Antiangiogenic agents in cancer chemotherapy. Totowa (NJ)7 Humana Press; 1999. p. 317 - 99. [14] Maeda K, Chung YS, Takatsuka S, et al. Tumor angiogenesis as a predictor of recurrence in gastric carcinoma. J Clin Oncol 1995;13: 477 - 81. [15] Bochner BH, Cote RJ, Weidner N, et al. Angiogenesis in bladder cancer: relationship between microvessel density and tumor prognosis. J Natl Cancer Inst 1995;87:1603 - 12. [16] Olivarez D, Ulbright T, DeRiese W, et al. Neovascularization in clinical stage A testicular germ cell tumor: prediction of metastatic disease. Cancer Res 1994;54:2800 - 2. [17] Moehler TM, Neben K, Ho AD, et al. Angiogenesis in hematologic malignancies. Ann Hematol 2001;80:695 - 705. [18] Lindmark G, Gerdin B, Sundberg C, et al. Prognostic significance of the microvascular count in colorectal cancer. J Clin Oncol 1996;14:461 - 6. [19] Takebayashi Y, Aklyama S, Yamada K, et al. Angiogenesis as an unfavourable prognostic factor in human colorectal carcinoma. Cancer 1996;78:226 - 31. [20] Busam KJ, Berwick M, Blessing K, et al. Tumor vascularity is not a prognostic factor for malignant melanoma of the skin. Am J Pathol 1995;147:1049 - 56. [21] Fallowfield ME, Cook MG. The vascularity of primary cutaneous melanoma. J Pathol 1991;164:241 - 4. [22] Akslen LA, LiVolsi VA. Increased angiogenesis in papillary thyroid carcinoma but lack of prognostic importance. Hum Pathol 2000;31: 439 - 42.
G. Valente et al. [23] Franquemont DW, Fechner RE, Mills SE. Histologic classification of sinonasal intestinal-type adenocarcinoma. Am J Surg Pathol 1991;15: 368 - 75. [24] Sauter ER, Nesbit M, Watson JC, et al. Vascular endothelial growth factor is a marker of tumor invasion and metastasis in squamous cell carcinomas of the head and neck. Clin Cancer Res 1999;5: 775 - 82. [25] Saaristo A, Partanen TA, Arola J, et al. Vascular endothelial growth factor–C and its receptor VEGFR-3 in the nasal mucosa and in nasopharyngeal tumors. Am J Pathol 2000;157:7 - 14. [26] Neuchrist C, Erovic BM, Handisurya A, et al. Vascular endothelial growth factor C and vascular endothelial growth factor receptor 3 expression in squamous cell carcinomas of the head and neck. Head Neck 2003;25:464 - 74. [27] O-charoenrat P, Rhys-Evans P, Eccles SA. Expression of vascular endothelial growth factor family members in head and neck squamous cell carcinoma correlates with lymph node metastasis. Cancer 2001; 92:556 - 68. [28] Ninck S, Reisser C, Dyckhoff G, et al. Expression profiles of angiogenic growth factors in squamous cell carcinomas of the head and neck. Int J Cancer 2003;106:34 - 44. [29] Tae K, El-Naggar AK, Yoo E, et al. Expression of vascular endothelial growth factor and microvessel density in head and neck tumorigenesis. Clin Cancer Res 2000;6:2821 - 8. [30] Hahnfeldt P, Panigrahy D, Folkman J, et al. Tumor development under angiogenic signaling: a dynamical theory of tumor growth, treatment response, and postvascular dormancy. Cancer Res 1999;59:4770 - 5. [31] Relf M, LeJeune S, Scott PA, et al. Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet-derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer Res 1997;57:963 - 9. [32] Weidner N. Tumour vascularity and proliferation: clear evidence of a close relationship. J Pathol 1999;189:297 - 9. [33] Bhattacharyya N. Cancer of the nasal cavity: survival and factors influencing prognosis. Arch Otolaryngol Head Neck Surg 2002; 128:1079 - 83. [34] Kamijo T, Yokose T, Hasebe T, et al. Potential role of microvessel density in predicting radiosensitivity of T1 and T2 stage laryngeal squamous cell carcinoma treated with radiotherapy. Clin Cancer Res 2000;6:3159 - 65. [35] Meert AP, Paesmans M, Martin B, et al. The role of microvessel density on the survival of patients with lung cancer: a systematic review of the literature with meta-analysis. Br J Cancer 2002;87: 694 - 701. [36] Wesseling P, van der Laak JA, Link M, et al. Quantitative analysis of microvascular changes in diffuse astrocytic neoplasms with increasing grade of malignancy. Hum Pathol 1998;29:352 - 8. [37] Beecken WD, Fernandez A, Joussen AM, et al. Effect of antiangiogenic therapy on slowly growing, poorly vascularized tumors in mice. J Natl Cancer Inst 2001;93:382 - 7.
www.elsevier.com/locate/humpath
Original contributions
Prognostic significance of microvessel density and vascular endothelial growth factor expression in sinonasal carcinomasB Guido Valente MDa,*, Carlo Mamo MDc, Antonella Bena MDc, Elisa Prudente BTechna, Cristina Cavaliere BScd, Simonetta Kerim MDd, Giuseppina Nicotra BScb, Alberto Comino MDe, Giorgio Palestro MDd, Ciro Isidoro MDb, Fabio Beatrice MDf a
Pathology Section, Department of Medical Sciences, Amedeo Avogadro University Medical School, I 28100 Novara, Italy Laboratory of Molecular Pathology, Department of Medical Sciences, Amedeo Avogadro University Medical School, I 28100 Novara, Italy c Epidemiology Unit, Agenzia Regionale Protezione Ambientale, I 10095 Grugliasco, Italy d Department of Biomedical Sciences and Human Oncology, School of Medicine and San Giovanni Hospital, I 10100 Torino, Italy e Laboratory of Anatomic Pathology, Santa Croce and Carle Hospital, I 12100 Cuneo, Italy f Department of Otolaryngology, Giovanni Bosco Hospital, I 10100 Torino, Italy b
Received 9 August 2005; revised 18 November 2005; accepted 21 November 2005
Keywords: Sinonasal carcinoma; Microvessel density; Tumor angiogenesis
Summary The prognostic significance of microvessel density and proliferative activity of the neoplastic cells, evaluated respectively by CD31 and Ki-67 positivity, and immunohistochemical expression of vascular endothelial growth factor (VEGF) was retrospectively investigated in 105 cases of sinonasal carcinoma (80 surgical specimens and 25 biopsies). The most represented histologic types were intestinal-type adenocarcinoma found in 36 patients (34.3%), squamous cell carcinoma (SCC) in 34 (32.4%), mucinous adenocarcinoma (mainly made up of signet-ring cell patterns) in 15 (14.3%), and adenoid cystic carcinoma in 7 (6.7%). Microvessel density values (in vessels per square millimeter), VEGF, and Ki-67 were not dependent on histologic type but were rather correlated to the histologic grading in SCC. Clinical data were available for 92 (87.6%) of 105 patients, with minimum follow-up of 48 months. Most of the patients (81.5%) were at an advanced stage (T3-T4) at diagnosis. The values of all markers were correlated to tumor stage ( P = .03). Multivariate analysis showed that both microvessel density and proliferative activity of the neoplastic cells were independent prognostic parameters (mortality hazard ratio, 1.33 and 1.60, respectively). Although VEGF expression was not correlated to prognosis on the whole series ( P = .06), it was a powerful prognostic marker when the analysis was restricted to the group of SCCs (hazard ratio, 3.02; 90% confidence interval, 1.58-5.80). These results show that tumor neoangiogenesis, expressed by microvessel density, together with proliferative activity, is a pathologic marker with a strong prognostic impact in sinonasal carcinomas. Therefore, it may be a
B
This work was supported by a grant from the Piedmont Region. * Corresponding author. E-mail address: [email protected] (G. Valente). 0046-8177/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2005.11.021
392
G. Valente et al. useful tool in this field so as to carry out therapeutic protocol planning, which may be further enhanced by the adoption of the more recent antiangiogenic molecules. D 2006 Elsevier Inc. All rights reserved.
1. Introduction When evaluating tumor growth, neoangiogenesis is known to be one of the essential events in neoplastic progression. A great deal of interest has been paid to the predictive value of neoangiogenesis (represented as microvessel density determined by the evaluation of CD31, CD34, and FVIIIr positivity) on the clinical progression and prognosis of several types of tumors over the last decade [1]. Indeed, it has been demonstrated that microvessel density has a prognostic role for a number of epithelial tumors, including carcinomas of the breast [2], lung [3], and large bowel [4], as well as for melanoma [5] and gliomas [6]. The regulation of the neoangiogenetic process is complex; normal tissues have a number of factors with proangiogenic or antiangiogenic function, produced by stromal and inflammatory cells, which are strictly balanced in normal tissues, whereas proangiogenic factors are usually prevalent in neoplastic tissues. Immunohistochemical and/or molecular techniques have been used to demonstrate a relationship between neoangiogenesis and expression of growth factors with proangiogenetic activity, such as vascular endothelial growth factor (VEGF) and its subunits [7]. The expression of these molecules has been seen to have a prognostic significance in many epithelial and nonepithelial tumors. This study analyzes the microvessel density and the growth fraction (by CD31 and Ki-67 positivity, respectively), as well as the immunohistochemical expression of VEGF, in a multicentric series of 105 cases of carcinomas that developed in the nose and paranasal sinuses, with the aim of assessing their prognostic role. These tumors are relatively infrequent and often related to occupational exposures, and to date, their prognosis has been based on traditional clinical and pathologic features. TNM staging and histologic type were seen to be particularly significant, as it has been observed that patients with T3 to T4 stages and squamous cell or undifferentiated carcinoma have a worse outcome, as also recently documented [8], whereas glandular tumors seem to be associated with a better prognosis. Moreover, localization and treatment modality also show an influence on the actuarial survival rate. Because diagnosis is frequently made at an advanced stage and the outcome remains ominous in many cases, as an average of only 40% of these patients have a 5-year survival after first diagnosis [8], a need to acquire biologic markers such as those analyzed in the present study clearly exists to estimate the aggressiveness of these tumors with as much accuracy as possible. Furthermore, data on the neoangiogenesis may also have therapeutic implications
when the latest reports about the application of antiangiogenic substances in solid tumors are taken into consideration.
2. Materials and methods 2.1. Patients The study was conducted on 105 cases of malignant epithelial tumors of nasal cavity and paranasal sinuses, mainly obtained from the files of the pathology departments of the University of Torino (San Giovanni Hospital) and Novara and from Cuneo Hospital (Italy). The sources of the other cases have been listed in the acknowledgment section. Table 1 Main clinicopathologic data of the 105 cases included in this study Sex Age (y)
Male Female 20-29 30-39 40-49 50-59 60-69 70-79 N80
Histologic type Intestinal-type adenocarcinoma Grade 1 Grade 2 Grade 3 SCC Grade 1 Grade 2 Grade 3 Mucinous adenocarcinoma Adenoid cystic carcinoma Oncocytoid adenocarcinoma Low-grade NOS adenocarcinoma Basaloid SCC Schneiderian carcinoma Verrucous carcinoma Others (low-grade mucoepidermoid carcinoma, small cell carcinoma, teratocarcinosarcoma) Surgical therapy Yes No Abbreviation. NOS, not otherwise specified.
n
%
84 21 2 8 18 47 19 9 2
80.0 20.0 1.9 7.6 17.1 44.8 18.1 8.6 1.9
36 4 23 9 34 0 13 21 15 7 2 2
34.3
2 2 2 3
1.9 1.9 1.9 2.8
80 25
76.2 23.8
32.4
14.3 6.7 1.9 1.9
Microvessel density and VEGF expression in carcinomas Table 2 Clinicopathologic findings in 92 cases with available clinical files Staging T1 T2 T3 T4 Surgical therapy Yes No Chemotherapy Yes No Radiotherapy Yes No Metastasis Yes No Relapse Yes No
n
%
1 16 22 53
1.1 17.4 23.9 57.6
74 18
80.4 19.6
22 70
23.9 76.1
80 12
86.9 13.1
7 85
7.6 92.4
41 51
44.6 55.4
The main clinicopathologic data are summarized in Table 1. The study cohort included 84 men and 21 women, with an average age of 64.3 F 11.1 years (range, 28-83 years). Surgical approach, followed by postoperative radiotherapy and/or chemotherapy, was the first choice therapy in 80 cases (76.2%). The remaining 25 patients were not operable and were treated exclusively by radiochemotherapy. All routine histologic slides were reviewed and histologic diagnoses were reformulated, according to the World Health Organization [9]: intestinal-type adenocarcinoma (36 cases) was the most represented histologic variety, followed by 34 squamous cell carcinomas (SCCs), 15 mucinous adenocarcinomas (with prevalence of signet-ring cell patterns), and 7 adenoid cystic carcinomas. More unusual tumors were recognized in the other 13 cases. The separation of intestinal-type adenocarcinomas and mucinous adenocarcinomas was justified by the worse prognosis demonstrated by the latter. The histologic grading was also reformulated, according to the World Health Organization criteria, for SCC [9]. Because of its low reproducibility, histologic grading of adenocarcinoma was not formulated. Both clinical files and radiologic documentation were available to help in the new determination of the clinical staging at first diagnosis in 92 patients (72 men and 20 women) (Table 2): 1 patient was reclassified as stage I, 16 patients (17.4%) as stage II, 22 (23.9%) as stage III, and 53 (57.6%) as stage IV. Follow-up ranged from 48 to 96 months.
2.2. Immunohistochemical studies The following monoclonal antibodies were used on paraffin sections: clone JC70A (DAKO A/S, Glostrup,
393 Denmark) to identify CD31, clone MIB-1 (DAKO A/S) to identify Ki-67, and clone C-1 (Santa Cruz Biotechnology, Santa Cruz, Calif) and clone 26503.11 (Sigma, St Louis, Mo) to identify human VEGF. The anti-VEGF monoclonal antibodies were raised against the 140 and 165 amino acids isoforms, respectively. The sections had previously been placed in 10 mmol/L citrate buffer, pH 6, and microwaved at 750 W for two 5-minute periods. Immunostaining was performed using the labeled streptavidin-biotin method (LSAB2 prediluted system kit, DAKO A/S), and diaminobenzidine 3,3-tetrachloride was used as chromogen. Negative controls omitted the primary antibody. The evaluation of CD31 positivity was restricted to microvessels inside the intratumoral stroma or on the front of invasion, according to the bhot spotQ method [10]. At first, 10 fields for each case showing a high microvessel density were selected at low magnification; the vessels were then counted at high magnification (40 objective) in each field. The square of the microscopic field was determined, and the average value was compared with 1 mm2 extension for each case. Three groups of cases were arbitrarily individualized on the basis of the results: less than 100, 100 to 150, and more than 150 vessels per square millimeter. Vascular endothelial growth factor positivity was displayed in the cytoplasms of the neoplastic cells with a more or less strong intensity and was evaluated by the following semiquantitative method: , negative or only sporadic positive cells; +, 5% to 20% positive cells; ++, 20% to 70% positive cells; and +++, more than 70% positive cells (+ was identified with the number 1 (+), 2 (++), and 3 (+++) for the statistical evaluation). The nuclei of the neoplastic cells characteristically showed Ki-67 positivity. Ten high-magnification fields were evaluated for each case, and the percentage of the neoplastic positive cells was calculated. Finally, the average percentage of the positive cells was obtained. Three groups of cases
Fig. 1 Intestinal-type adenocarcinoma. CD31-positive neoformed vascular structures are particularly numerous on the front of invasion (immunoenzymatic method with streptavidin, original magnification 275).
394
G. Valente et al.
2.3. Specificity of the antihuman VEGF antibody The specificity of the anti-VEGF monoclonal antibody used in the present study was ascertained by using Western blotting of VEGF expression in cultured human colorectal cancer DLD1 cells. Cell homogenate was prepared by freeze-thawing cycles and sonication in a buffer containing 2% sodium deoxycholate and a cocktail of protease inhibitors. Fifty micrograms of protein was heat-denatured in a Laemmli buffer not containing b-mercaptoethanol, separated by sodium dodecyl sulfate–polyacrilamide gel electrophoresis, and then electroblotted onto nitrocellulose filters. Western blotting was performed after a standard procedure as described elsewhere [11] using the anti-VEGF as primary antibody and a peroxidase-conjugated rabbit antimouse IgG polyclonal antiserum as secondary antibody.
2.4. Statistical analysis Fig. 2 Squamous cell carcinoma. CD31-positive neoformed vessels are well represented close to neoplastic sheets (immunoenzymatic method with streptavidin, original magnification 275).
were arbitrarily individualized: less than 25% positive neoplastic cells, 25% to 50% positive cells, and more than 50% positive cells. All immunohistochemical reactions were evaluated by 2 persons (G. V. and S. K.) for each case; the interobserver difference was less than 10%. In a limited number of cases, sections were also processed for immunofluorescence staining. In this case, deparaffinized biopsies were incubated overnight at 48C with the primary antibody (diluted 1:50 in phosphate buffer containing 0.1% Triton X-100 and 4% fetal calf serum). After having washed out the excess of the antibody, sections were subsequently incubated for 1 hour at room temperature with a fluorescein isothiocyanate–conjugated goat antimouse polyclonal antiserum as the secondary antibody (diluted 1:350, as aforementioned). All incubations were performed in a humidified chamber protected from light. Negative control was performed by omitting the primary antibody. Sections were counterstained with propidium iodide (1:500 with the secondary antibody) to label the nucleus. Table 3
The v 2 statistic method was used to check the association between 2 variables (for instance microvessel density and tumor stage). Survival curves were estimated by the product-limit (Kaplan-Meier) method. The log-rank test was performed to compare 2 or more survival curves for different groups of individuals. Cox proportional hazards models were used to investigate associations of microvessel density, VEGF expression, and growth factor with diseasefree survival. Analysis was carried out by SAS statistical software (SAS Institute Inc, Cary, NC).
3. Results 3.1. CD31 immunostaining and microvessel density CD31 was displayed by endothelial cells as well as by a fraction of the tumor-infiltrating mononuclear cells. Vascularization was mainly localized in the stroma surrounding the neoplastic tissue and was prominent in the invasive front (Figs. 1 and 2). Microvessel counts varied from 50 to 232.95/mm2, with a median value of 121.8/mm2 (average, 123.9/mm2). The most represented histologic types had values ranging from 50.6 to 209.8 in intestinal-type adenocarcinomas (median value, 122.72), whereas they were
Positivity for CD31, VEGF, and growth fraction in the most frequent histologic types of sinonasal carcinoma Microvessel density (CD31) (/mm2)
Intestinal-type adenocarcinoma SCC Mucinous adenocarcinoma Adenoid cystic carcinoma
VEGF
Growth fraction (% positive cells)
b100
100-150
1 (+)
2 (++)
3 (+++)
b25
25-50
15 (41.7)
13 (36.1)
8 (22.2)
13 (36.2)
7 (19.4)
16 (44.4)
12 (33.3)
17 (47.2)
7 (19.5)
11 (32.4) 6 (40.0)
13 (38.2) 7 (46.7)
10 (29.4) 2 (13.3)
13 (38.2) 7 (46.7)
7 (20.6) 3 (20.0)
14 (41.2) 5 (33.3)
2 (5.9) 6 (40.0)
15 (44.1) 5 (33.3)
17 (50.0) 4 (26.7)
1 (14.3)
4 (57.1)
2 (28.6)
1 (14.3)
4 (57.1)
2 (28.6)
5 (71.4)
2 (28.6)
Values are presented as n (%).
N150
N50
0 (0)
Microvessel density and VEGF expression in carcinomas Table 4
Positivity for CD31, VEGF, growth fraction, and clinical stage in 31 cases of intestinal-type adenocarcinoma Microvessel density (CD31) (/mm2)
T1 T2 T3 T4
(0) (6) (11) (14)
395
VEGF
Growth fraction (% positive cells)
b100
100-150
N150
1 (+)
2 (++)
3 (+++)
b25
25-50
N50
/ 5 (83.3) 3 (27.3) 4 (28.6)
/ 1 (16.7) 5 (45.4) 6 (42.8)
/ / 3 (27.3) 4 (28.6)
/ 2 (33.4) 4 (36.4) 5 (35.7)
/ 2 (33.3) 2 (18.2) 2 (14.3)
/ 2 (33.3) 5 (45.4) 7 (50.0)
/ / 5 (45.4) 4 (28.6)
/ 3 (50.0) 3 (27.3) 7 (50.0)
/ 3 (50.0) 3 (27.3) 3 (21.4)
Values are presented as n (%). Slash (/) represents a zero value.
from 59.1 to 172.72 (median value, 106.81) in mucinous and signet-ring cell adenocarcinomas, from 50.0 to 224.24 (median value, 119.86) in SCC, and from 58.63 to 187.87 (median value, 125.0) in adenoid cystic carcinomas (Table 3). Although microvessel density did not show any significant correlation with the histologic type of the tumors, the v 2 test demonstrated a significant association between microvessel density values and the clinical stage ( P = .001) (Tables 4 and 5) and with poor histologic grading in SCC ( P = .04).
3.2. Vascular endothelial growth factor immunostaining The presence and localization of VEGF in biopsy sections of tumor specimen were analyzed using immunocytochemical methods. In Fig. 3, an example of immunofluorescence staining is shown; strong positive granules are localized in the cytoplasms of some neoplastic cells in a SCC. Extensive analysis of VEGF expression in all tumor specimens was performed by immunohistochemical staining using the streptavidin-biotin peroxidase method as previously described. No substantial differences in localization and intensity of the reaction product were observed between the 2 antibodies used, although clone C-1 showed a slightly higher background staining. Vascular endothelial growth factor positivity was almost completely restricted to the neoplastic tissue and was expressed in the cytoplasm of neoplastic cells in all cases (Figs. 4 and 5). There was a rather uniform distribution of the positive neoplastic cells in the tumor tissue. The semiquantitative evaluation showed a high number of positive cells (N75%) in 16 (44.4%) of 36 cases of intestinal-type adenocarcinomas, in 14 (41.1%) of 34 cases of SCC, and in 5 (33.3%) of 15 mucinous and signet-ring cell adenocarcinomas (Table 3). More than 50% of the neoplastic cells expressed VEGF in 6 of 7 adenoid cystic carcinoma (85.7%). The expression of VEGF was associated with microvessel Table 5
(0) (6) (7) (17)
3.3. Specificity of the anti-VEGF Specificity of the anti-VEGF antibody used in the present work was ascertained by Western blotting analysis of VEGF expression in cell homogenate of human colorectal cancer DLD1 cells. As shown in Fig. 6, the antibody revealed the presence of bands corresponding to the polypeptides of expected size for VEGF isoforms [12].
3.4. Growth fraction The growth fraction was evidenced by the nuclear positivity with Ki-67 antibody (Figs. 7 and 8). High levels of proliferative activity were observed in SCC, as more than 50% of Ki-67–positive neoplastic cells were counted in 17 (50%) of 34 cases, whereas similar values were observed in only 7 (19.4%) of 36 intestinal-type adenocarcinomas and 4 (26.6%) of 15 mucinous and signet-ring cell adenocarcinomas (Table 3). The growth fraction was also seen to be associated with clinical stage ( P = .03) (Tables 4 and 5) and histologic grading in SCC ( P b .0001).
3.5. Patients outcome Patients with high microvessel density (N150 vessels per square millimeter) showed a worse 5-year disease-free survival than those with intermediate (100-150/mm2) or low (b100/mm2) microvessel density (Fig. 9). These survival curves differed significantly even when adjusted for the effect of the histologic type (log-rank test, P = .03). Multivariate analysis revealed that microvessel density was a significant independent prognostic factor; mortality hazard ratio adjusted for sex, age, therapy, staging, and histologic
Positivity for CD31, VEGF, growth fraction, and clinical stage in 30 cases of SCC Microvessel density (CD31) (/mm2)
T1 T2 T3 T4
density ( P = .01) and with poor histologic grading in SCC ( P = .04). It was, moreover, associated with clinical stage ( P = .03) (Tables 4 and 5).
VEGF
Growth fraction (% positive cells)
b100
100-150
N150
1 (+)
2 (++)
3 (+++)
b25
25-50
N50
/ 5 (83.3) 1 (14.4) 3 (17.7)
/ 1 (16.7) 3 (42.8) 9 (52.9)
/ / 3 (42.8) 5 (29.4)
/ 2 (33.3) 3 (42.8) 7 (41.2)
/ 1 (16.7) / 5 (29.4)
/ 3 (50.0) 4 (57.2) 5 (29.4)
/ / 1 (14.4) 1 (5.9)
/ 3 (50.0) 2 (28.4) 9 (52.9)
/ 3 (50.0) 4 (57.2) 7 (41.2)
Values are presented as n (%). Slash (/) represents a zero value.
396
G. Valente et al.
Fig. 5 Squamous cell carcinoma. Neoplastic cells displaying intense VEGF positivity (immunoenzymatic method with streptavidin, original magnification 275). Fig. 3 Squamous cell carcinoma. Immunofluorescence staining with monoclonal antibody against human VEGF revealed with fluorescein isothiocyanate–conjugated secondary antibody (green staining). Nuclei are stained with propidium iodide (red staining). Vascular endothelial growth factor positivity is displayed by the cytoplasm of neoplastic cells (original magnification 200).
staging, and histologic type was 1.60 (90% CI, 1.06-2.42) for increasing level of Ki-67. Vascular endothelial growth factor positivity, when expressed semiquantitatively, was not related to the survival on the overall series. However, this marker was seen to have a high impact on the disease-free survival for the patients with SCC (log-rank test, P = .07) (Fig. 11).
type was 1.33 (90% confidence interval [CI], 0.90-1.95) for increasing level of microvessel density. When Ki-67 was categorized on 3 levels (b25%, 25%-50%, and N50% positive cells), patients with a high growth fraction (N50% of positive neoplastic cells) also showed a worse 5-year disease-free survival (log-rank test, P = .15) (Fig. 10). However, the prognosis was worse in the first group (b25%) than in the second (25%-50%) at 5-year follow-up. Ki-67 positivity was a significant independent prognostic factor; mortality hazard ratio adjusted for sex, age, therapy,
Fig. 4 Mucinous adenocarcinoma. Strong VEGF positivity in the cytoplasms of the neoplastic cells (immunoenzymatic method with streptavidin, original magnification 275).
Fig. 6 Specificity of antihuman VEGF antibody used in the present study. Vascular endothelial growth factor peptides were identified by Western blotting analysis. Molecular weights correspond to the expected size for dimeric forms of VEGF peptides made up of 121, 165, 189, and 206 amino acids. The 2 anti-VEGF monoclonal antibodies gave similar results.
Microvessel density and VEGF expression in carcinomas
397
4. Discussion
in agreement with previous observations on other tumors, that is, microvessel density, as measured by CD31 immunostaining, is a highly predictive and independent marker of both progression and prognosis. Neoangiogenesis is correlated to other progression parameters, mainly the clinical stage. Furthermore, the entity of the neoangiogenesis is not related to the intrinsic aggressiveness of the different histologic types because low microvessel densities were also evaluated in varieties known to have a worse prognosis, such as the signet-ring cell adenocarcinoma [23]. This may explain the powerful impact of microvessel density in our histologically heterogeneous series. There was a correlation between microvessel density and the immunohistochemical expression of VEGF (the most investigated of proangiogenic proteins); a similar result was observed in many other experimental and human tumors, for instance, in head and neck carcinomas [22,24], because the neovascularization should be directly up-regulated by VEGF and other proangiogenic factors. Moreover, in normal conditions, also VEGF seems to be implicated in the development of blood [25] and lymphatic vessels through its receptor, VEGFR3 (flt-4) [26]. Indeed, this is an interesting observation because lymphatic diffusion is prominent in head and neck tumors. Vascular endothelial growth factor proved to be a progression marker for SCC in our series, as demonstrated in other head and neck sites [27]. High amounts of VEGF have been found in supernatants of head and neck SCC [28], leading to the hypothesis of a more consistent role in these, rather than in other histologic types. However, apart from the predictive value found in the SCC group, the expression of VEGF lacked a significant prognostic value in our whole series as in other reports [29]. The discrepancy between this observation and the strong prognostic role of microvessel density, in spite of their relationship, may be explained by the fact that the expression of single
When evaluated by the hot spot method, microvessel density proved to be a valuable prognostic indicator in several human tumors. The validation of neovascularization as a relevant prognostic tool has been widely confirmed in most retrospective studies on breast carcinoma, non–small cell lung carcinoma, and prostate carcinoma, which showed that microvessel density is directly correlated with the clinical outcome in these tumors [13]. Positive results have also been obtained in several studies on gastric [14] and bladder carcinoma [15] and testicular germinal cell tumors [16], as well as hematologic malignancies [17]. Contradictory results on prognosis have been found in colorectal carcinomas [18,19] and in malignant melanoma of the skin [20,21], whereas in some tumors such as papillary thyroid carcinoma, tumor angiogenesis was correlated to the development or maintenance of the tumor but not to the outcome [22]. There is a shortage of data on the predictive value of microvessel density in sinonasal carcinomas. Our results are
Fig. 8 Squamous cell carcinoma. Neoplastic sheets are made up by cells with marked Ki-67 positivity (immunoenzymatic method with streptavidin, original magnification 275).
Fig. 7 Intestinal-type adenocarcinoma with high growth fraction. Most of the neoplastic cells show nuclear Ki-67 positivity (immunoenzymatic method with streptavidin, original magnification 275).
Multivariate analysis revealed that VEGF had a weak prognostic impact on the whole series; hazard ratio adjusted for sex, age, surgical therapy, staging, and histologic type was 1.10 (90% CI, 0.79-1.53) for increasing level of VEGF. Noteworthy was the fact that VEGF was seen to be an important prognostic factor in SCC; hazard ratio was 3.02 (90% CI, 1.58-5.80) in these patients because of the increase in VEGF level when adjusted for sex, age, therapy, staging, and histologic grading.
398
G. Valente et al.
Fig. 9 Kaplan-Meier survival curves by level of microvessel density (3 levels) in 92 cases of malignant epithelial tumors of nasal cavity show that survival is significantly worse in patients with microvessel density higher than 150/mm2. The differences between the curves has been evaluated by log-rank test ( P = .03).
proangiogenic or antiangiogenic factors is a dynamic, more than a static, biologic process, balanced by both positive and negative angiogenesis regulators [30], and may vary over time [31]. Analysis of the growth fraction, as determined by MIB-1 immunohistochemical staining, had a significant and independent prognostic role in sinonasal carcinomas. To the best of our knowledge at the time of writing, no relevant results have been reported as to what impact proliferative
activity has on these tumors. A direct relationship between proliferative activity and neoangiogenesis was expected because a good vascular support is necessary to warrant the growth of the neoplastic cells, particularly on the front of invasion [32]. The growth fraction represents a further strong prognostic evaluation element from a practical point of view. Correlations between biologic parameters and clinical outcome have rarely been reported in nasal tumors, and the
Fig. 10 Kaplan-Meier survival curves by level of growth fraction (3 levels of Ki-67 positivity). The 5-year disease-free survival is significantly worse in patients with more than 50% of positive neoplastic cells. After 3 years, patients with low growth fraction (b25%) show a worse prognosis than those with intermediate value (25%-50%). The differences between the curves has been evaluated by log-rank test ( P = .15).
Microvessel density and VEGF expression in carcinomas
399
Fig. 11 Survival curves restricted to cases of SCC demonstrate that VEGF expression is a powerful marker of prognosis in these tumors. Cases belonging to group 3 (N75% positive cells) show a negative 5-year disease-free survival. The differences between the curves has been evaluated by log-rank test ( P = .07).
prognosis of these tumors, also in the most recently reported series, has mainly been formulated on the basis of traditional parameters [8,33]. As expected, these later proved to be prognostically significant also in our series. However, the clinical application of our results could be of great relevance, in as much as the diagnosis of many tumors of nose and paranasal sinuses is often formulated too late to allow a surgical approach and/or alternative therapies such as exclusive radiotherapy that are able to reduce the neoplastic mass. Microvessel density has been demonstrated to be a useful tool in the prediction of radiosensitivity in laryngeal SCC [34], and antiangiogenic therapies used to reduce a neoplastic mass that cannot be operated on and are now applied in several clinical trials. Whether microvessel density may be an indicator of the susceptibility to the treatment with antiangiogenic molecules or alternatively a predictive marker of response is still debatable. Even if microvessel density is used to reflect the angiogenesis, the two parameters cannot be considered to completely overlap [35]. Moreover, although tumors with high microvessel density should be, from a rational point of view, the most suitable candidates for antiangiogenic therapy [36], there is experimental evidence that, in mice, less vascularized bladder tumors give the best response [37]. In conclusion, sinonasal carcinomas can be added to the many other malignant tumors where prognosis is related to the neoangiogenesis. Both microvessel density and growth fraction have emerged as being powerful prognostic and predictive tools in our study. These results could well indicate a further application in the light of therapeutic approach using antiangiogenic molecules that are potentially useful in these neoplasias that are often diagnosed at
an advanced stage for which an extrasurgical approach is preferred.
Acknowledgments The authors thank the following pathologists who contributed to the article by sending their cases: Dr Enrico Bollito (San Luigi Hospital, Orbassano, Italy), Dr Bruno Torchio and Dr Manuela Motta (Mauriziano Hospital, Torino, Italy), Dr Stefano Taraglio and Dr Laura Davico Bonino (Maria Vittoria Hospital, Torino, Italy), Dr Mauro Giudici (Ospedale degli Infermi, Biella, Italy), Dr Clemente Vineis (Ospedale Civile di Ivrea, Italy), Dr Mario Abrate (ASL 17 of Savigliano, Fossano e Saluzzo, Italy), and Dr Giovanni Cera (ASL 16 di Mondovı`, Italy). The authors thank Mrs Barbara Wade for revising the English language and Dr. Gabriella Nicosia and Mr Denis Longhi for the skillful technical assistance.
References [1] Hasan J, Byers R, Jayson GC. Intra-tumoural microvessel density in human solid tumours. Br J Cancer 2002;86:1566 - 77. [2] Gasparini G, Toi M, Verderio P, et al. Prognostic significance of p53, angiogenesis, and other conventional features in operable breast cancer: subanalysis in node-positive and node-negative patients. Int J Oncol 1998;12:1117 - 25. [3] Giatromanolaki A, Koukourakis MI, Kakolyris S, et al. Vascular endothelial growth factor, wild-type p53, and angiogenesis in early operable non–small cell lung cancer. Clin Cancer Res 1998;4:3017 - 24. [4] Banner BF, Whitehouse R, Baker SP, et al. Tumor angiogenesis in stage II colorectal carcinoma: association with survival. Am J Clin Pathol 1998;109:733 - 7.
400 [5] Neitzel LT, Neitzel CD, Magee KL, et al. Angiogenesis correlates with metastasis in melanoma. Ann Surg Oncol 1999;6:70 - 4. [6] Schmidt NO, Westphal M, Hagel C, et al. Levels of vascular endothelial growth factor, hepatocyte growth factor/scatter factor and basic fibroblast growth factor in human gliomas and their relation to angiogenesis. Int J Cancer 1999;84:10 - 8. [7] Hlatky L, Hahnfeldt P, Folkman J. Clinical application of antiangiogenic therapy: microvessel density, what it does and doesn’t tell us. J Natl Cancer Inst 2002;94:883 - 93. [8] Dulguerov P, Jacobsen MS, Allal AS, et al. Nasal and paranasal sinus carcinoma: are we making progress? A series of 220 patients and a systematic review. Cancer 2001;92:3012 - 29. [9] Shanmugaratnam K, Sobin LH, editors. Histological typing of tumours of the upper respiratory tract and ear. 2nd ed. WHO international histological classification of tumours. Berlin7 SpringerVerlag; 1991. p. 32. [10] Weidner N, Semple JP, Welch WR, et al. Tumor angiogenesis and metastasis—correlation in invasive breast carcinoma. N Engl J Med 1991;324:1 - 8. [11] Isidoro C, De Stefanis D, Demoz M, et al. Expression and posttranslational fate of Cathepsin D in HT-29 tumor cells depend on their enterocytic differentiation state. Cell Growth Differ 1997;8:1029 - 37. [12] Ferrara N, Henzel WJ. Pituitary follicular cells secret a novel heparinbinding growth factor specific for vascular endothelial cells. Biochem Biophys Res Commun 1989;161:851 - 8. [13] Gasparini G, Harris AL. Prognostic significance of tumor vascularity. In: Teicher BA, editor. Antiangiogenic agents in cancer chemotherapy. Totowa (NJ)7 Humana Press; 1999. p. 317 - 99. [14] Maeda K, Chung YS, Takatsuka S, et al. Tumor angiogenesis as a predictor of recurrence in gastric carcinoma. J Clin Oncol 1995;13: 477 - 81. [15] Bochner BH, Cote RJ, Weidner N, et al. Angiogenesis in bladder cancer: relationship between microvessel density and tumor prognosis. J Natl Cancer Inst 1995;87:1603 - 12. [16] Olivarez D, Ulbright T, DeRiese W, et al. Neovascularization in clinical stage A testicular germ cell tumor: prediction of metastatic disease. Cancer Res 1994;54:2800 - 2. [17] Moehler TM, Neben K, Ho AD, et al. Angiogenesis in hematologic malignancies. Ann Hematol 2001;80:695 - 705. [18] Lindmark G, Gerdin B, Sundberg C, et al. Prognostic significance of the microvascular count in colorectal cancer. J Clin Oncol 1996;14:461 - 6. [19] Takebayashi Y, Aklyama S, Yamada K, et al. Angiogenesis as an unfavourable prognostic factor in human colorectal carcinoma. Cancer 1996;78:226 - 31. [20] Busam KJ, Berwick M, Blessing K, et al. Tumor vascularity is not a prognostic factor for malignant melanoma of the skin. Am J Pathol 1995;147:1049 - 56. [21] Fallowfield ME, Cook MG. The vascularity of primary cutaneous melanoma. J Pathol 1991;164:241 - 4. [22] Akslen LA, LiVolsi VA. Increased angiogenesis in papillary thyroid carcinoma but lack of prognostic importance. Hum Pathol 2000;31: 439 - 42.
G. Valente et al. [23] Franquemont DW, Fechner RE, Mills SE. Histologic classification of sinonasal intestinal-type adenocarcinoma. Am J Surg Pathol 1991;15: 368 - 75. [24] Sauter ER, Nesbit M, Watson JC, et al. Vascular endothelial growth factor is a marker of tumor invasion and metastasis in squamous cell carcinomas of the head and neck. Clin Cancer Res 1999;5: 775 - 82. [25] Saaristo A, Partanen TA, Arola J, et al. Vascular endothelial growth factor–C and its receptor VEGFR-3 in the nasal mucosa and in nasopharyngeal tumors. Am J Pathol 2000;157:7 - 14. [26] Neuchrist C, Erovic BM, Handisurya A, et al. Vascular endothelial growth factor C and vascular endothelial growth factor receptor 3 expression in squamous cell carcinomas of the head and neck. Head Neck 2003;25:464 - 74. [27] O-charoenrat P, Rhys-Evans P, Eccles SA. Expression of vascular endothelial growth factor family members in head and neck squamous cell carcinoma correlates with lymph node metastasis. Cancer 2001; 92:556 - 68. [28] Ninck S, Reisser C, Dyckhoff G, et al. Expression profiles of angiogenic growth factors in squamous cell carcinomas of the head and neck. Int J Cancer 2003;106:34 - 44. [29] Tae K, El-Naggar AK, Yoo E, et al. Expression of vascular endothelial growth factor and microvessel density in head and neck tumorigenesis. Clin Cancer Res 2000;6:2821 - 8. [30] Hahnfeldt P, Panigrahy D, Folkman J, et al. Tumor development under angiogenic signaling: a dynamical theory of tumor growth, treatment response, and postvascular dormancy. Cancer Res 1999;59:4770 - 5. [31] Relf M, LeJeune S, Scott PA, et al. Expression of the angiogenic factors vascular endothelial cell growth factor, acidic and basic fibroblast growth factor, tumor growth factor beta-1, platelet-derived endothelial cell growth factor, placenta growth factor, and pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer Res 1997;57:963 - 9. [32] Weidner N. Tumour vascularity and proliferation: clear evidence of a close relationship. J Pathol 1999;189:297 - 9. [33] Bhattacharyya N. Cancer of the nasal cavity: survival and factors influencing prognosis. Arch Otolaryngol Head Neck Surg 2002; 128:1079 - 83. [34] Kamijo T, Yokose T, Hasebe T, et al. Potential role of microvessel density in predicting radiosensitivity of T1 and T2 stage laryngeal squamous cell carcinoma treated with radiotherapy. Clin Cancer Res 2000;6:3159 - 65. [35] Meert AP, Paesmans M, Martin B, et al. The role of microvessel density on the survival of patients with lung cancer: a systematic review of the literature with meta-analysis. Br J Cancer 2002;87: 694 - 701. [36] Wesseling P, van der Laak JA, Link M, et al. Quantitative analysis of microvascular changes in diffuse astrocytic neoplasms with increasing grade of malignancy. Hum Pathol 1998;29:352 - 8. [37] Beecken WD, Fernandez A, Joussen AM, et al. Effect of antiangiogenic therapy on slowly growing, poorly vascularized tumors in mice. J Natl Cancer Inst 2001;93:382 - 7.