Abstract
BACKGROUND:
No clear guidelines exist for management of breast cancer brain metastases (BCBM).
OBJECTIVE:
We assessed the relationship between patient and tumor characteristics, treatment, and overall survival (OS).
METHODS:
We conducted a retrospective review of 196 patients who received brain radiation for BCBM between 2009–2013 at Mayo Clinic. Primary tumor characteristics were collected, including simplified molecular subtype. Other characteristics included patient’s ECOG, number of brain lesions at BCBM diagnosis, and treatment received, including neurosurgery, whole-brain radiation therapy (WBRT), and stereotactic radiosurgery (SRS). The primary endpoint was OS from time of BCBM diagnosis.
RESULTS:
Single-variable analysis revealed patients with HER2+ breast cancer had improved OS (
CONCLUSIONS:
Patients with the best OS had an ECOG of 0, HER2+ disease, and 1–3 brain lesions. The best OS was associated with the combination of neurosurgery and radiation therapy. A comprehensive treatment plan including neurosurgical evaluation and radiation therapy should be considered for patients with BCBM.
Introduction
The incidence of brain metastases in patients diagnosed with breast cancer has been poorly recorded, with the best historical estimates ranging between 10 and 30% of patients who have metastatic disease [25]. In recent years, the detection rate of brain metastases in patients living with cancer appears to have increased because of multiple factors, including improved imaging modalities and increased survival time from initial cancer diagnosis [14]. For example, in patients with human epidermal growth factor receptor 2 positive (HER2+ ) tumors, significant improvements in overall survival (OS) have been accomplished with the development of trastuzumab (Herceptin® ; F. Hoffmann-La Roche, Basel, Switzerland); however, trastuzumab is known to have poor penetration of the blood brain barrier, creating a reservoir for tumors to grow and increasing the risk of isolated brain metastases in this patient population [6].
Despite significant strides in the treatment of systemic breast cancer over the last few decades, the prognosis of breast cancer brain metastasis (BCBM) remains poor, with median survival rates of approximately 6 months; however, more recent small retrospective studies have shown an increase in the median OS, approximately 14 months for patients treated with multimodality therapy and certain subtypes [12,24,27,28,30,33]. Current therapeutic treatment options for patients diagnosed with brain metastases include whole-brain radiation treatment (WBRT), stereotactic radiosurgery (SRS), neurosurgery, or chemotherapy. These may be used alone or in combination, but none are without significant adverse events. Recent data comparing WBRT and SRS suggests that patients with a single brain metastasis may have improved survival from the addition of SRS to WBRT, but it does not appear to affect survival in patients with multiple metastases [1]. Similarly, some studies indicate that WBRT in combination with neurosurgery is superior to WBRT alone in patients with a single brain metastasis [34]. However, in this study and others, only a limited number of breast cancer patients have been included. There is limited evidence in this area to suggest which patients should be treated and the appropriate regimens to use.
Breast cancer and brain metastases are widely heterogeneous diseases, and as such, they should not be approached in the same manner every time. Prognostic indicators such as a patient’s functional status, hormone and HER2 receptor status, the number of brain lesions, and the extent of systemic disease should be taken into consideration when reviewing treatment options. Currently, clinicians may subjectively base therapy decisions on these factors, but there are no generalized standard practice guidelines for the treatment of patients with newly-diagnosed BCBM. A recent multidisciplinary expert panel reviewed the available literature specific to HER2+ BCBM and published a consensus guideline with recommendations based on patient prognosis and extent of disease; however, they noted that there was not sufficient data available to provide a traditional evidence-based clinical practice guideline, and their recommendations were limited to patients who had HER2+ disease [36].
The goal of this article is to review the importance of individual patient and tumor characteristics for determining prognosis from time of BCBM diagnosis, evaluate the role of neurosurgery and radiation therapy, and potentially contribute to the future development of a standardized approach to the treatment of patients who have been newly diagnosed with BCBM.
Patients and methods
Patients
After receiving IRB approval we reviewed female patients with breast cancer treated at Mayo Clinic between January 2009 and December 2013. Mayo Clinic sites in Minnesota, Florida, and Arizona were included. Eligible patients had to have brain metastases visible on computed tomography or magnetic resonance imaging and had to have received radiation therapy to the brain between January 2009 and December 2013. The patients’ electronic medical records had to be available for review with all data needed for this study.
Methods
Data was extracted retrospectively from the patients’ medical records. Data points collected included the date of primary breast cancer diagnosis and the date of BCBM diagnosis. The primary breast tumor characteristics were collected, including simplified molecular subtype: hormone receptor positive (HR+ ), HER2+ , and triple negative (TN). The Eastern Cooperative Oncology Group (ECOG) performance status, rated 0–5, was recorded from the time of BCBM diagnosis [31]. The number of brain lesions or the presence of only leptomeningeal disease was recorded on the basis of imaging. Data was collected on the treatment received, including neurosurgical removal of the BCBM lesions and the type of radiation administered: either WBRT, SRS, or the combination of WBRT and SRS. The date of death and whether it was breast cancer-related were also recorded.
Statistical analyses
Continuous variables were summarized with the sample median and range. Categorical variables were summarized with number and percentage. The Kaplan-Meier curve analysis and the log-rank test were used to assess OS following a diagnosis of BCBM. Single-variable Cox proportional hazards regression models were used to evaluate exploratory associations between the patients’ pathological, clinical, or surgical characteristics and OS following the diagnosis of BCBM; hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated. No adjustment for multiple testing was made in these exploratory analyses, and
Results
Patient characteristics
A summary of the pathological, clinical, and surgical information from the 196 study patients is displayed in Table 1.
Table 2 summarizes simplified molecular subtype, number of brain lesions, ECOG at BCBM diagnosis, treatment received including neurosurgery, WBRT and SRS, and the association with the primary endpoint OS. All patients with HER2+ disease were treated with trastuzumab.
Simplified molecular subtype
OS was different among the breast cancer subtypes (Fig. 1). With HR+ disease as the reference group, patients with HER2+ disease had the best OS (16.55 months,
Number of brain lesions
Patients with 1–3 brain lesions had the best median OS (12.58 months) and were used as the reference group. Patients with leptomeningeal disease, diagnosed by magnetic resonance imaging and lumbar puncture with cerebral spinal fluid cytology, had a median OS of 6.13 months and a 2.52-fold higher risk of death compared to patients with 1–3 brain lesions (
ECOG status
The ECOG score at the time of BCBM diagnosis was associated with OS. For every 1 point increase in ECOG score the OS worsened by approximately 4 months. Patients who had an ECOG score of 0 had the longest OS of 14.19 months compared with a median OS of 10.5 months for patients with an ECOG score of 1 (HR = 1.53,
Radiation therapy
All patients received WBRT, SRS, or a combination of WBRT and SRS. The majority of patients (44%) received WBRT alone, with dose ranges from 2000 to 3000 cGy, and this was used as the reference (Figure 4). There was no significant difference in OS between WBRT alone and SRS alone (
Surgery & radiation therapy
Figure 4 shows the different treatment modalities, including neurosurgical resection, WBRT, SRS, or WBRT and SRS. The patients receiving only WBRT and no neurosurgery were used as the reference group and had a median OS of 6.08 months. There was a significant decrease in the risk of death when neurosurgery was added to WBRT, resulting in a longer median OS of 8.53 months (
Discussion
Differences in the incidence and survival of BCBM are reflective of the different primary breast cancer subtypes. Patients with primary HR+ BCBM have historically had the best OS, and patients with TN BCBM have had the worst OS [12,16,30]. This is especially true during the pre-trastuzumab treatment era. In more recent retrospective studies, including ours, patients with primary HER2+ breast cancer had the best median OS by a significant margin (16.5 months) [21,28]. Additionally, patients with HER2+ BCBM have a higher incidence of BCBM than patients with HR+ or TN breast cancer [4,13,29]. Although the reasons why patients with HER2+ BCBM survive longer are still unknown, they may be impacted by better systemic disease control with trastuzumab. In a study by Kirsch et al., patients with HER2+ disease and BCBM who were not treated with trastuzumab had a similar OS to HR+ patients [20,26].
In addition to simplified molecular subtype, the number of brain lesions correlated with survival. Patients with only leptomeningeal involvement had the worst OS. This is consistent with the poor prognosis previously seen in these patients, including a retrospective series of 68 women who had breast cancer with isolated leptomeningeal involvement that reported a median OS of 4 months and a 1-year survival rate of 13% [23]. As the number of brain lesions increased, survival also worsened.
In general, poor performance status has consistently been associated with OS from brain metastases from any cancer [22,39]. The performance status has been assessed in various studies utilizing different scales, including the ECOG score, Karnofsky performance status, and the Graded Prognostic Assessment [18,38]. In our study, the ECOG score was recorded for each patient at the time of BCBM diagnosis. With each 1-point increase in ECOG score, which correlates with worsening overall performance status, there was significantly worse OS. These findings are consistent with previous studies of BCBM [17,24,40]. This prognostic information is critical to developing an accurate discussion with patients and their families. With this information, patients and physicians can select appropriate treatment options, including palliative or hospice care.
The mainstay of treatment for BCBM is radiation therapy. The most common radiation treatment utilized is WBRT [8]. With advancements in radiation therapy, there has been an increased use of SRS. Previous studies have provided conflicting results regarding the comparison of WBRT, SRS, and the combination of WBRT and SRS. Similar to previous reports, there was no difference in OS for our patients treated with WBRT or SRS alone [15]. A randomized, controlled trial of patients who had different cancer subtypes and brain metastases showed that compared with SRS alone, the use of WBRT combined with SRS did not improve OS for patients with 1–4 brain metastases [3]. However, the rate of intracranial relapse occurred more frequently in those who did not receive WBRT. Another study, with a primary focus on neurocognitive decline, found that, in patients with brain metastases from any cancer, 73% did not have central nervous system recurrence 1 year after SRS plus WBRT compared with 27% who only received SRS alone (
Recently, the randomized clinical trial NCCTG N0574 was presented at the American Society of Clinical Oncology 2015 annual meeting. The study was designed to primarily evaluate cognitive decline after SRS or WBRT combined with SRS treatment of brain metastases in patients will all types of primary malignancies [5]. Brown, et al. showed significant improvement in intracranial disease progression with the combination of SRS and WBRT compared to SRS alone, but there was no significant difference in median OS [5]. Additionally, combined therapy posed an increased risk of cognitive decline, which could impact quality of life [5]. This prospective study shows improved local control of brain metastases with SRS plus WBRT, but unfortunately, this study was limited in that it included patients with a spectrum of primary malignancies. There may be some relation to our data for breast cancer patients because of the importance of good intracranial disease control. For our patients, it is possible that the local control with WBRT plus SRS was an important factor in the significantly longer OS that was seen compared to patients receiving WBRT or SRS alone.
In our study, patients who received multimodality therapy including neurosurgery, WBRT, and SRS had the longest median OS of 23.16 months. This is consistent with previous reports favoring multimodality therapy [10,32]. The patient population that will benefit from this multimodal therapy remains controversial [11]. Patients with a solitary metastatic brain lesion may benefit from neurosurgery or SRS [2], but the addition of WBRT could potentially cause detrimental neurocognitive effects, and the data do not support that this addition can slow BCBM progression [19,35]. For now, our recommendation would be that neurosurgery or SRS should be pursued first followed by WBRT for patients with limited disease (1–3 lesions). If neurosurgery is used first, salvage SRS can be considered for recurrence [9]. In patients with good systemic disease control, performance status consideration can be given to combined treatment with neurosurgery, SRS, and WBRT to provide the best OS [10]. For patients with 4 or more BCBM, we would recommend WBRT initially. For patients with an ECOG score of 3–4, we would recommend the best supportive care.
Our study has some limitations that will be further evaluated in a prospective fashion. These limitations include the use of systemic and intrathecal chemotherapy, which could have influenced patient outcomes. Although most BCBM cases have the same simplified molecular subtype as the primary breast cancer, this cannot be confirmed in the patients who did not undergo BCBM resection. Along with TN and HER2+ breast cancer subtypes, several other prognostic factors have been identified for the development of BCBM, including younger age, ethnicity, and the presence of BRCA1 germ-line mutations [37,41]. These factors can be explored in a future retrospective analysis. It would also be interesting to assess the extent of extracranial disease in future analyses as this may influence treatment decisions. Our results of describing the clinical outcomes after using a comprehensive treatment plan that included neurosurgical evaluation and the addition of WBRT and SRS therapy based on clinicopathological characteristics can be used for patient counseling and to design prospective trials evaluating new modalities of therapy.
Footnotes
Acknowledgements
This work had no specific funding. Accepted for Poster Presentation at the American Society for Clinical Oncology 2015 Annual Meeting: Abstract #1095.
Conflicts of interest
Dr. Perez is an employee of Genentech, Inc. (employment began after the completion of this work). All other authors declare no conflicts of interest.