ANTICANCER RESEARCH 27: 2839-2844 (2007)
Prognostic Value of HER2 and Progesterone Receptor Expression in Endometrial Carcinoma with Positive Peritoneal Washing MARIA BENEVOLO1, AMINA VOCATURO1, FLAVIA NOVELLI1, LUCIANO MARIANI2, GIUSEPPE VOCATURO2, ANNA MARIA CIANCIULLI3, FERDINANDO MARANDINO1, RAFFAELE PERRONE-DONNORSO1, DIANA GIANNARELLI4, PIER GIORGIO NATALI5 and MARCELLA MOTTOLESE1
Departments of 1Pathology, 2Gynecologic Oncology and 3Clinical Pathology, Unit and 5Immunology Laboratory, Regina Elena Cancer Institute, 00144, Rome, Italy
4Biostatistic
Abstract. Background: Although the majority of endometrial cancer (EC) patients can be cured by surgery, unexpected recurrent disease may also occur in early stage patients. In the present study, whether or not the analysis of multiple biopathological parameters might lead to more accurate predictions of the clinical outcome of EC patients with longterm follow-up (FU) was investigated. Patients and Methods: Estrogen and progesterone receptor (ER and PgR) positivity and HER2 overexpression by immunohistochemistry were evaluated. The peritoneal washings (PWs) were analyzed by cytology and immunocytochemistry employing AR-3 and B72.3 monoclonal antibodies. Results: The patients with positive PW and HER2 positive tumors showed shorter overall survival compared to those bearing HER2 negative tumors (p=0.004). HER2 overexpression also influenced the patient outcome in the group with tumors lacking PgR (p=0.004). At multivariate analysis PgR and HER2 overexpression emerged as independent prognostic factors. Conclusion: The combined analysis of these biopathological markers could provide useful information for the selection of patients to be enrolled in innovative therapeutic strategies. Endometrial carcinoma (EC) is the most common invasive malignant tumor of the female genital tract and its occurrence has risen worldwide. Although as many as 70% of these patients can be cured by surgery, often in combination with radiotherapy, unexpected recurrent disease may also occur in patients with EC limited to the uterus (1). There are
Correspondence to: Marcella Mottolese, Ph.D., Regina Elena Cancer Institute, Pathology Department, via Elio Chianesi 53, 00144, Rome, Italy. Tel: +39 06 52666139, Fax: +39 06 52666139, e-mail:
[email protected] Key Words: Uterine neoplasms, erbB2, steroid receptors, peritoneal cytology, prognosis.
0250-7005/2007 $2.00+.40
many pathological features which appear to correlate with prognosis. These include cervical extension, depth of myometrial invasion, histotype and differentiation grade and also malignant peritoneal microdissemination, which is recognised as an adverse prognostic factor mainly in those patients with no extensive myometrial invasion or macroscopic omental and/or peritoneal implants (2). In this context, the use of a selected panel of monoclonal antibodies (MoAbs) (3) may increase the diagnostic accuracy of peritoneal cytology, significantly improving the intraoperative staging of EC and enhancing the chance of a cure. EC is generally known to be an endocrine-related neoplasm and loss of progesterone receptors (PgR) is associated with a more aggressive behaviour of the tumor which, consequently, no longer responds to specific hormonal treatments (4). Furthermore, PgR appears to be one of the most reliable independent parameters for predicting survival (5). In recent years, the molecular analysis of EC has identified abnormalities in the expression, structure, or activity of oncogene products which can contribute to the development and maintenance of the malignant phenotype (6). In particular HER2 overexpression and/or amplification have been widely investigated, although at present its prognostic value is still cause for debate (7, 8). Moreover, very few data are currently available regarding the prognostic impact of HER2 status concerning PgR expression or peritoneal cytology. In order to address this issue, HER2 was evaluated together with PgR in a large series of EC with known peritoneal washing (PW) cytological diagnosis and with longterm follow-up (FU).
Patients and Methods Patients. Written informed consent was obtained from 200 EC patients who were then admitted to the Regina Elena Cancer Institute (Rome, Italy) between 1988 and 2000. Primary surgery, consisting of Piver’s II type, was performed carrying out pelvic lymph-node sampling in patients with deep myometrial invasion
2839
ANTICANCER RESEARCH 27: 2839-2844 (2007) and macroscopic endocervical involvement. The patients age ranged from 36 to 84 years (median 62 years), 90.6% of the women were of postmenopausal and 9.4% of premenopausal status. According to the current surgical staging system (9), 135 patients were classified as stage Ia to Ic, 22 as stage II, 43 as stage IIIa to IIIc. Patients with advanced disease (stage IVa-b) were excluded from the study. Our series of patients included 178 endometrioid and 22 papillary serous adenocarcinomas. Mixed and uncommon histotypes were excluded from this study. The myometrial invasion, determined in 198 patients, was less than one half in 146 patients and more than one half in 52. A total of 71 tumors were well differentiated (G1), while 78 were moderately differentiated (G2) and 51 poorly differentiated (G3) (10). Sampling and preparation of peritoneal washings. Intraoperative PW was performed in 181 patients. Physiological solution (250 ml) at 37ÆC was injected into the peritoneal cavity and collected using heparin as anticoagulant. The cells, separated by centrifugation, were resuspended in phosphate buffer (pH 7.2) at a density of 1.106 cells/ml. The erythrocytes were eliminated by treating the PW with Tris NH4Cl (pH 7.4) for 10 minutes at 37ÆC. Multiple cytospins, obtained with a cytocentrifuge (Shandon Runcorn, Cheshire, UK), were stained following the Papanicolaou method, or fixed for 10 minutes in cold absolute acetone for immunocytochemistry (ICC) evaluation. Immunocytochemistry (ICC), immunohistochemistry (IHC) and fluorescence in situ hybridization (FISH). The MoAbs used for ICC of the PWs, were B72.3 (Menarini, Florence, Italy) and AR-3 (kindly provided by Dr. M. Prat, Turin University, Italy). The two MoAbs, used in combination, are able to recognize more than 95% of EC (11, 12). The PWs were considered positive when isolated cells or clusters of cells were found to be ICC reactive with at least one of the MoAbs employed in two separate cytospins. HER2 overexpression was determined by IHC in 5 Ìm paraffin sections using MoAb 300G9 (13) and to confirm the results obtained, the slides were also stained with MoAb CB11 (Menarini). A 96.8% concordance between the two reagents was demonstrated. No antigen retrieval treatment was used for MoAb 300G9, whereas pretreatment in citrate buffer, pH6, was performed for MoAb CB11. The HER2 immunoreactivity, evaluated using a 0 to 3+ scale, was considered positive when at least 10% of neoplastic cells displayed a moderate (score 2+) or strong (score 3+) staining of the entire plasma membrane. The tumors displaying weak (score 1+) or no staining (score 0) of the plasma membrane were considered negative. The estrogen receptor (ER; clone 6F11, Biogenex, Space, Milan, Italy) and PgR (clone 1A6, Biogenex) status was evaluated on paraffin embedded tissues. Before staining, the slides were pretreated as previously described. The cut-off for both receptors was 10% positive nulei. Immunostaining was performed using a streptavidin-biotin immunoperoxidase method (LSAB2 kit; Dako, Milan, Italy) and the enzymatic activity was developed using 3amino-9-ethyl-carbazole (AEC, Dako) as chromogenic substrate. FISH assay was performed according to the manufacturer’s recommendations (PathVysion Kit; Vysis, Downers Grove, IL, USA). The chromosome enumeration probe 17 (CEP17) was used as a control to determine the copy number of chromosome 17 to adjust for the effect of aneuploid chromosome 17 when the HER2 copy numbers were counted. Samples with a ratio value ≥2 were considered to be amplified.
2840
Statistical analysis. Associations between cytologic and ICC diagnosis were investigated by using Pearsons’ Chi-square test. For the purpose of our study, overall survival (OS) was considered as a measure of poor outcome (14). The survival curves were estimated using the Kaplan-Meier method. The log-rank test was used to assess differences between curves. Significance was defined at the p
