Subsequent to obtaining approval from the institutional review board of Qingdao University School of Medicine Ethics Committee, 259 patients with early stage breast cancer were recruited to the study between January 2010 to January 2011 at the Department of Breast Surgery, Qingdao Central Hospital, the Second Affiliated Hospital of Qingdao University (Qingdao, China). All patients provided written informed consent prior to their inclusion in the study. Information about the patients' clinical history was obtained from the patient records database of Qingdao Central Hospital.

The age of each patient was defined at diagnosis. The size of the primary tumor was considered to be the largest tumor diameter reported by a pathologist subsequent to surgical excision. Patients with early stage breast cancer were staged as I and II based on the tumor-node-metastases (TNM) classification of the International Union against Cancer, revised in 2010 (12). The histological grade of the tumors was classified according to the World Health Organization criteria (13). The lymph node status was determined by the presence of histological evidence for lymph node metastasis. All patients in the present study were treated with breast conserving surgery with or without axillary lymph node dissection. Following surgery, the patients received standard whole breast irradiation at the radiation oncology facilities of the Department of Therapeutic Radiology, Qingdao Central Hospital, the Second Affiliated Hospital of Qingdao University. The median total dose was 50 Gy to the whole breast. Adjuvant systemic chemotherapy and/or adjuvant hormone therapy were administered as clinically indicated in accordance with standard practice during the time of the study. Local recurrence-free survival (RFS), distant metastasis-free survival (DFS) and overall survival (OS) were the endpoints evaluated.

Formalin-fixed, paraffin-embedded tissues from the Department of Pathology, Qingdao Central Hospital were used in the construction of the breast cancer sample TMA. Briefly, representative areas of breast tumor samples were selected from whole tissue sections by pathologists. A cylindrical specimen core of 2-mm diameter was extracted from these regions and precisely arrayed into a new recipient paraffin block using a custom-built precision instrument (Beecher Instruments; Estigen OÜ, Tartu, Estonia) as previously described (14). The resulting TMA blocks were cut to 4-µm thickness with a microtome and placed on slides with an adhesive tape-transfer method (InstruMedics; Stryker Corporation, Kalamazoo, MI, USA). One section from each TMA was stained with hematoxylin and eosin (in hematoxylin for up to 2 min and eosin for 30 sec at room temperature) to verify the presence of tumor cells. The TMA was assessed for the staining intensity, and the percentage of stained cells in the nucleus, cytoplasm or membrane.

Immunohistochemical analysis was performed on the 4-µm thick paraffin-embedded tissue sections of the TMA using a standard streptavidin-peroxidase method. The subsequent levels of tissue sections were stained with Ventana Benchmark XT Autostainer according to the manufacturer's instructions (Ventana Medical Systems, Inc., Tucson, AZ, USA). The slides were incubated with a rabbit anti-human primary antibody against POSTN (dilution 1:500, cat. no. sc-67233; Santa Cruz Biotechnology, Inc., Dallas, TX, USA), ER (cat. no. G07268), PR (cat. no. Y12992), HER2 (cat. no. Y08422) and Ki-67 (cat. no. Y13569) (all from Roche Diagnostics, Indianapolis, IN, USA) for 28–32 min at 37°C. The primary antibodies for ER, PR, HER2, Ki-67 and the goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (Ultraview Universal DAB detection kit; cat. no. 05269806001; Roche Diagnostics) were ready-to-use for 8 min at 37°C. A known positive case was stained as a positive control. The primary antibody was replaced with nonimmune mouse serum (Santa Cruz Biotechnology, Inc.) as a negative control. Immunohistochemical staining was developed in a Dako Autostainer Plus using an LSAB detection kit (Dako; Agilent Technologies, Inc., Santa Clara, CA, USA). The individuals performing immunohistochemical evaluation were blinded to the clinical information. The staining results for all markers were first determined independently, then reviewed together at a multi-head microscope. According to the criteria for immunohistochemical evaluation reported previously (10), the presence of 1% of positively stained neoplastic cells was defined as positive POSTN expression.

Statistical analysis was performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA). The association of POSTN expression with categorical clinicopathological parameters was assessed by χ² tests. Estimates of RFS, DFS and OS were calculated by the Kaplan-Meier product-limit method, and the differences were assessed by the log-rank test. Multivariate analysis was performed using the Cox proportional hazards regression model to determine the prognostic effect on the independent contribution of each variable to survival; hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) were calculated. All Р-values are two-sided; Р<0.05 was considered to indicate a statistically significant difference.

The intracellular localization of POSTN protein staining was examined, as its location may affect the mechanism of action and potentially, the response to treatment. In the breast cancer tissue samples, there was a heterogeneous distribution of POSTN in the cytoplasm. Nuclear POSTN expression was relatively infrequent. Representative POSTN immunostaining results are included in Fig. 1.

The clinical and pathological variables of the patients are included in Table I. The median age at diagnosis for all patients was 53 years (range, 27–93), with 44.4% of the patients (n=115) <50 years at the time of diagnosis. A total of 45% (n=117) of the patients presented with lymph node metastases (N1 or 2) at the time of surgery. Of the 259 patients, 27% had the Luminal A molecular subtype, 33% Luminal B, 13% HER2-overexpressed and 25% triple-negative; the remaining patients were uncategorized.

As of February 2015, the median follow-up time was 51 months. Of all the patients in the study, 7% (n=17) of patients experienced local relapse, 12% (n=32) experienced distant metastasis and 11% (n=29) succumbed to the disease.

Clinicopathological characteristics and molecular features were compared with the POSTN expression status for the patients in the cohort of the present study. The results are summarized in Table I. The POSTN status was significantly associated with the histological grade (P<0.01), nodal status (P<0.05), molecular subtype (P<0.01), estrogen receptor (ER) status (P<0.01), progesterone receptor (PR) status (P<0.01) and Ki-67 expression (P<0.05). No significant association between POSTN expression and other parameters could be established.

Five-year survival analysis was performed for the 259 patients, considering the RFS, DFS and OS rates. As summarized in Table II and illustrated in Fig. 2, tumors with positive POSTN expression were associated with poorer outcomes for all endpoints, including RFS (P<0.05), DFS (P<0.01) and OS (P<0.01).

Multivariate analysis with the Cox proportional hazards model was performed using the following variables: POSTN expression, age, tumor size, histological grade, nodal status, disease subtype, ER, PR and HER2 status, and Ki-67 expression (Table III). Positive POSTN expression was associated with a relatively poor outcome for all endpoints. A larger tumor size, the tumor subtype and high Ki-67 expression were associated with reduced RFS time (P=0.007, P=0.005 and P=0.012, respectively). Positive nodal status was also associated with a reduced DFS and OS time (P=0.005 and P=0.003, respectively). A higher histological grade and high Ki-67 expression were further associated with a reduced OS time (P=0.043 and P=0.016, respectively).

Therefore, it was demonstrated by the multivariate analysis that POSTN expression retained significance independent of other prognostic factors, including as an independent prognostic marker for local relapse, distant metastasis and OS in the patients with early stage breast cancer of the present study.