We conducted a retrospective analysis using the medical records of patients who had been diagnosed with brain metastasis (BM) originating from breast cancer between January 2003 and January 2009. BM was diagnosed via magnetic resonance imaging (MRI) or computed tomography (CT) and/or lumbar puncture in cases in which patients evidenced symptoms such as headache, dizziness, nausea, or neurological deficits. This study was approved by the institutional review board of the Dong-A University Medical Center.

Patients with BM were treated as indicated with at least one of the following treatment modalities: whole brain radiotherapy (WBRT), surgical resection, and systemic treatments including chemotherapy and endocrine therapy. Corticosteroid treatment was administered after diagnosis, and continued throughout the ensuing treatment regimen. All patients with brain metastases were treated WBRT. The total dose was 30 Gy, applied in 10 fractions over two weeks. Intrathecal therapy for patients with leptomeningeal disease consisted of lumbar puncture or Ommaya reservoir with methotrexate 10-15 mg twice a week until malignant cells could no longer be detected in the cerebrospinal fluid.

Immunohistochemical analyses of ER and PR status were conducted using 4-µm sections of paraffin-embedded tissue specimens that were stained with monoclonal antibodies for ER (DAKO Co., Carpinteria, USA) and PR (DAKO Co.). We chose a cut-off level for negative cases of less than 10% positive nuclei with weak staining intensity, which is equal to Allred score 2. The tumor was considered to be HER2-positive if the primary or metastatic tumor was scored 3+ by HER2 IHC (Neo Markers, Fremont, USA), or in the case of IHC 2+ with HER2 gene amplification identified via the FISH method. TNBCs were defined as those that were ER-negative, PR-negative, and HER2-negative.

Chi-square and Fisher's exact tests were utilized as categorical variables to compare patient characteristics among the three time period groups. A Kruskall-Wallis test was conducted in order to compare the median values of nonparametric variables. We established independent variables as follows: age at the time of diagnosis, tumor stage at primary diagnosis of breast cancer, tumor grading, vessel invasion, adjuvant systemic treatment, and palliative systemic treatment.

Disease-free survival was calculated from the date of initial diagnosis to the date of first recurrence. Overall survival was calculated from the date of diagnosis of BM to the date of death or final follow-up. The interval between initial diagnosis and BM was calculated as the time from the date of initial diagnosis to the date at which BM was radiologically proven. The interval between the distant metastasis and BM was calculated as the time elapsed from the detection of the first recurrence to the radiological evidence of BM in breast cancer patients.

Survival durations were calculated via the Kaplan-Meier method. The log-rank test was employed for the comparison of cumulative survival in the patient groups. To identify independent factors significantly related to patient prognosis we used Cox's proportional hazard analysis with a stepwise procedure. In all tests, p<0.05 was established as the threshold of statistical significance. The SPSS software program (version 14.0; SPSS Inc., Chicago, USA) was employed in all analyses.

The clinical characteristics of the BM patients and the relationship with receptor status are presented in Table 1. Twenty four patients (40.0%) were diagnosed with TNBC. The median age at diagnosis of BM in the TNBC groups was 49 yr (range, 30-74 yr). Three patients had stage I disease at diagnosis, 9 patients had stage II, 12 patients had stage III at primary presentation. Approximately 88% of patients had T1 or T2 disease, and 70% were lymph node positive. Thirteen patients had histologic grade (HG) 1 or 2 tumors, and the others had HG 3 tumors. The most frequent presenting signs and symptoms of BM were headache, nausea, and vomiting, followed by dizziness, double vision, and unilateral weakness of the extremities. The brain was the first site of metastasis in 7 patients (29.2%) and was accompanied, in the majority of cases, by other sites of metastatic disease. The most common sites of extracranial metastasis were the lung (60%), bone (50%), liver (45%), and the lymph nodes (45%). In 6 patients (25.0%), a single BM was detected, and in 18 patients multiple parenchymal BM occurred; leptomeningeal carcinomatosis was not observed in any of the TNBC patients. However, the clinicopathologic parameters did not differ according to the breast cancer subtype (Table 1).

The disease-free survival durations of non-TNBC and TNBC patients were 25.6 months and 17.9 months, respectively (p=0.135). The median time intervals from distant metastasis to BM were 19.5 months and 8.4 months, respectively (p=0.006). The median times from initial diagnoses to BM were 43.7 and 25.5 months, respectively (p=0.027; Figure 1). Primary tumor characteristics and the interval duration from initial diagnosis to BM are provided in Table 2. Histologic grade (p=0.036), ER status (p=0.005), vascular invasion (p=0.007), subtype (p=0.027) were significantly related to the time interval from initial diagnosis to BM. However, only histologic grade was independent prognosis factor (relative risk [RR], 0.039; 95% confidence interval [CI], 1.034-3.668; p=0.039). Among 17 HER2-positive patients, 5 received trastuzumsab treatment. These patients tended to exhibit a prolonged interval from initial diagnosis to BM (p=0.057; Table 2).

The median follow-up after the diagnosis of CNS metastasis was 21 months (range, 0.7-96 months), and the median overall survival from the date of BM was 7.6 months (95% CI, 5.1-10.1 months). At 1 yr, 20% of the patients remained alive, whereas 10% were alive at 2 yr. As most patients evidenced active disease outside the brain, it was difficult to determine whether the patients died of intracranial or extracranial metastatic disease.

Table 3 shows overall survival according to selected patient characteristics. Patients suffering from single BM evidence longer overall survival durations (19.3 months) as compared with patients who had multiple BM (7.6 months; p=0.021). Patients who responded to WBRT tended to have longer overall survival durations (p=0.025). Median survival was 14.6 months in patients who received palliative chemotherapy versus 1.4 months for those who did not (p<0.001). We analyzed 17 HER2-positive patients according to trastuzumab therapy. No significant difference in median survival were noted after a diagnosis of BM between the tratuzumab-treated patients and the untreated patients (8.5 months vs. 7.3 months, respectively, p=0.579). And, overall survival from brain metastasis to death among non-TNBC and TNBC patients was significantly different (7.6 months and 4.3 months, respectively, p=0.046; Figure 2). A multivariate analysis using the Cox-regression test identified two risk factors: no additional systemic treatment (RR, 2.714; 95% CI, 1.446-5.095; p=0.006) and TNBC (RR, 2.212; 95% CI, 1.162-4.209; p=0.016). However, number of brain metastases (p=0.050), and response to WBRT (p=0.387) were not independent prognostic factor for overall survival after BM.