We thank Nancy Pinnell and Barbara Silver of the Department of Radiation Oncology, Dana-Farber/Brigham and Women's Cancer Center for their help in the preparation of the manuscript.
Salvage stereotactic radiosurgery (SRS) is often considered in breast cancer patients previously treated for brain metastases. The goal of this study was to analyze clinical outcomes and prognostic factors for survival in the salvage setting.
The authors retrospectively examined 79 consecutive breast cancer patients who received salvage SRS (interval of >3 months after initial therapy), 76 of whom (96%) received prior whole-brain radiation therapy. Overall survival (OS) and central nervous system (CNS) progression-free survival rates were calculated from the date of SRS using the Kaplan-Meier method. Prognostic factors were evaluated using the Cox proportional hazards model.
Median age was 50.5 years. Fifty-eight percent of this population was estrogen receptor positive, 62% was HER2 positive, and 10% was triple negative. At the time of SRS, 95% had extracranial metastases, with 81% of extracranial metastases at other visceral sites (lung/pleura/liver). Forty-eight percent had stable extracranial disease. Median interval from initial brain metastases therapy to SRS was 8.4 months. Median CNS progression-free survival after SRS was 5.7 months (interquartile range [IQR], 3.6-11 months), and median OS was 9.8 months (IQR, 3.8-18 months). Eighty-two percent of evaluable patients received further systemic therapy after SRS. HER2 status (adjusted hazard ratio [HR], 2.4; P = .008) and extracranial disease status (adjusted HR, 2.7; P = .004) were significant prognostic factors for survival on multivariate analysis.
With advances in systemic therapies for breast cancer, central nervous system (CNS) relapse has emerged as an important source of morbidity and mortality. There has been increasing recognition that a predilection to CNS dissemination is influenced by breast cancer subtype, with HER2-positive and triple-negative tumors most likely to develop brain metastases.1-6 Furthermore, tumor subtype has emerged as a prognostic factor for survival in breast cancer patients with newly diagnosed brain metastases.7-11
Selection of patients with brain metastases for treatment with stereotactic radiosurgery (SRS) and the timing of treatment are controversial. SRS may be used in the upfront setting without whole-brain radiation therapy (WBRT) in patients with a limited number of metastases; 2 randomized studies were unable to demonstrate a survival benefit with the addition of WBRT.12, 13 An SRS boost after WBRT improves overall survival in patients with a single brain metastasis.14 SRS may also be used in the salvage setting at the time of CNS recurrence. Data regarding the outcomes of breast cancer patients who undergo SRS in the salvage setting are lacking, and studies have not addressed prognostic factors at the time of CNS recurrence. The objectives of this study are to report clinical outcomes in breast cancer patients treated with SRS at the time of recurrence after initial treatment with WBRT and to analyze for prognostic factors, including tumor subtype.
MATERIALS AND METHODS
This retrospective study was approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board. Records of all breast cancer patients who underwent SRS at Brigham and Women's Hospital between January 1, 2000 and June 30, 2010 were reviewed. As a priori intention-to-treat could not be assigned retrospectively, salvage SRS was defined as patients who underwent SRS at an interval of >3 months from the date of initial treatment for brain metastases (WBRT or other). Salvage SRS was further defined by 2 conditions: first, there was no stated intent in the medical records to deliver a radiosurgical boost upfront, and second, all patients had to have evidence of persistent or progressive CNS disease, defined as any growth of an existing enhancing lesion or the development of a new contrast-enhancing lesion. Furthermore, clinical notes from patients receiving SRS before the 3 month cutpoint were screened for stated intent of salvage to avoid misclassification. During the period of this study, institutional practice was to deliver upfront WBRT in breast cancer patients with multiple newly diagnosed brain metastases. The follow-up schedule after SRS was not uniform in all cases; however, the standard approach was to obtain brain magnetic resonance imaging (MRI) 2 months after SRS and approximately every 3 months thereafter.
This study was conducted with the approval of the Dana-Farber/Harvard Cancer Center Institutional Review Board. Review of the Radiation Oncology SRS database identified 132 consecutive patients with a histologic diagnosis of breast cancer who underwent SRS between January 1, 2000 and June 30, 2010. Follow-up data, including medical records and neuroimaging, were available for 130 patients. One patient was excluded who developed a brain metastasis secondary to a metachronous nonsmall cell lung cancer. Of the 129 evaluable patients, 79 met the above criteria for salvage SRS and constituted the study population.
Stereotactic Radiosurgical Technique
SRS was performed using the Novalis linear accelerator-based radiosurgery platform. Before 2009, all patients were immobilized with the use of a fixed head frame. As of April 2009, conventional frame-based radiosurgery at our institution was replaced by frameless delivery using the thermoplastic BrainLAB cranial mask immobilization system. Diagnostic MRI and contrast-enhanced simulation computed tomography (CT) datasets were coregistered for contouring purposes using BrainLAB iPlan. Iterative forward planning was performed using the BrainLAB BrainSCAN stereotactic treatment planning software. Prescribed dose was not standardized in this study. Generally, arcs using circular collimators, arcs using static multileaf collimators, or multiple static fields using dynamic MLCs were used for treatment. Institutional practice aims for 99% target coverage, that is, 99% of the target covered by the prescribed dose.
Recurrence was defined as either local or distant. Subsequent recurrence events after the first recurrence event were not recorded. Local recurrence at the site of SRS was defined as an increase in the diameter of the contrast-enhancing lesion of at least 25% on follow-up MRI.15 Distant recurrence was defined as the presence of a new enhancing lesion consistent with a brain metastasis or leptomeningeal enhancement outside of the SRS target volume.
The initial analysis included descriptive statistics of clinical and demographic information. Estimates of CNS recurrence-free survival and overall survival were calculated using the Kaplan-Meier method. In the absence of radiographic evidence of CNS recurrence patients were censored at the date of the last MRI demonstrating no evidence of recurrence. The effect of clinical and demographic covariates on survival was estimated using a Cox proportional hazards model. Variables with a P value <.1 on univariate analysis were used to construct a multivariate model. All statistical tests were 2-sided and used a significance level of .05. All analyses were performed using R (version 2.11.1) with R Commander and Survival packages.16-18
The following potential prognostic factors were analyzed: patient age at the date of SRS (dichotomized at 65 years as per the Radiation Therapy Oncology Group Recursive Partitioning Analysis); Karnofsky performance status (KPS) at the time of SRS (≤70% vs ≥80%); the number of metastases treated with SRS (1 vs 2-3 vs ≥4); the interval between initial treatment for brain metastases and SRS; the status of extracranial disease (stable vs progressing), as defined by any evidence of radiographic progression on restaging CT scanning ±2 months from the date of the SRS; estrogen receptor (ER) status; HER2 status; and triple-negative status. Tumors were assigned ER and HER2 status on the basis of immunohistochemistry (IHC) carried out on the primary tumor at the time of original diagnosis. Tumors were considered HER2 positive only if they scored 3+ by IHC or if they were 2+ by IHC and also amplified (ratio >2.0) on the basis of fluorescence in situ hybridization.
Of the 79 patients who underwent salvage SRS, 96% (76 of 79) received WBRT as part of the initial therapy for brain metastases (Table 1). Eighteen percent (14 of 79) underwent upfront surgical resection. The median time from initial treatment for brain metastases to SRS was 8.4 months (interquartile range [IQR], 4.9-15 months). The median age was 50.5 years (IQR, 44.1-57.7 years). Sixty-one percent of patients had 4 or more lesions at diagnosis of brain metastasis. Ninety-five percent had evidence of extracranial sites of metastatic disease at the time of SRS, with 81% of the study population having other visceral metastases (lung/pleura/liver). Only 48% had stable systemic disease at the time of SRS. However, this was a selected population of patients with good performance status. Ninety-seven percent had a KPS ≥70%. Data regarding the use of systemic therapy after SRS were incomplete. However, of the 62 patients for whom this information was available, 51 (82%) received further systemic therapy after SRS.
Table 1. Patient Characteristics (N=79)
Abbreviations: CNS, central nervous system; ER, estrogen receptor; KPS, Karnofsky performance status; PR, progesterone receptor; SRS, stereotactic radiosurgery; TDI, Total dural irradiation using photon/electron technique; WBRT, whole brain radiation therapy.
Type of initial therapy
WBRT + surgery
No. tumors treated by SRS
N of initial CNS metastases
Systemic disease status
In total, 291 tumors were treated by SRS in 79 patients. The median tumor volume treated was 0.55 mL (IQR, 0.16-2.0 mL). Median prescription dose was 18 Gray (Gy) (IQR, 16-20 Gy), with a median normalization of 78% (IQR, 70%-80%).
With a median follow-up of 16.7 months, 61% of patients (48 of 79) experienced progressive CNS disease. Sixty-six percent of recurrences were distant to the site of SRS (n = 32), 16% were local-only recurrences (n = 8), and 16% were simultaneous local and distant recurrences (n = 8). The median CNS recurrence-free survival time after SRS was 5.7 months (IQR, 3.6-11 months). Of the 48 patients who developed recurrent disease within the brain after SRS, 42 received further salvage therapy (SRS or stereotactic radiation therapy, n = 25; repeat WBRT, n = 5; posterior fossa radiation, n = 1; craniotomy, n = 5; or a trial of CNS-directed systemic therapy, n = 6).
Survival and Prognostic Factors
Across all patients, median survival time was 9.8 months (IQR, 3.8-18 months) after salvage SRS (Fig. 1). On univariate analysis, HER2-negative patients had a significantly increased risk of death, with an unadjusted hazard ratio (HR) of 2.3 (95% confidence interval [CI], 1.5-3.9; P = .002) (Fig. 2). On univariate analysis, other factors significantly associated with an increased risk of death included progressive extracranial disease (HR, 2.9; 95% CI, 1.7-4.9; P < .001) (Fig. 3), KPS ≤70% (HR, 4.2; 95% CI, 2.3-7.7; P < .001), and triple-negative status (HR, 3.8; 95% CI, 1.7-8.3; P = .001). Nonsignificant factors on univariate analysis included patient age, ER status, number of lesions treated with SRS, and the interval between initial treatment for brain metastases and SRS.
On multivariate analysis, HER2 negativity (HR, 2.4; 95% CI, 1.3-4.7; P = .008) and progressive extracranial disease (HR, 2.7; 95% CI, 1.4-5.2; P = .004) were found to be associated with increased risk of death after salvage SRS (Table 2, Fig. 4).
Table 2. Univariate and Multivariate Analyses for Factors Associated With Overall Survival
With advances in modern systemic therapy, an increasing number of patients are experiencing recurrent CNS progression events for which CNS-directed treatment modalities are being considered. Quality of life and the prevention of neurologic deficit are important goals of care in this population. SRS is a single-visit outpatient procedure that is associated with minimal acute toxicity. Shaw et al reported rates of irreversible grade 3 or higher acute toxicity of 0% to 7% in the setting of SRS for recurrent brain tumors.19 Therefore, it is an attractive treatment option to both patients and oncologists alike, particularly in the era of frameless radiosurgical delivery. However, the role of salvage SRS in the management of breast cancer patients who have recurrent brain metastases after initial therapy is unclear, and potential prognostic factors that might help guide decision making in this setting are undefined. Furthermore, given increasing concerns about cost in the management of patients with advanced cancer, an understanding of survival outcomes takes on additional importance.
Our study demonstrates protracted survival after SRS at the time of CNS progression for a selected population of breast cancer patients with good performance status, the large majority of whom received further systemic therapy after SRS. Our study suggests that HER2-positive disease and stable extracranial metastatic disease at the time of SRS are favorable prognostic factors for overall survival. Strengths of our study include that it reflects a modern series of patients treated in the trastuzumab era with a focus solely on the question of SRS use in the salvage setting. Other strengths include the completeness of the immunohistochemistry data and internal pathology review, and the availability of follow-up imaging.
There are several limitations inherent to our study. This is a retrospective single-institution experience, and as such is subject to selection bias. Of note, however, our study population was enriched with patients with multiple brain metastases and extracranial involvement, which a priori might have been expected to confer a worse prognosis. The use of systemic therapy before and after SRS was not well characterized in this study. Previous reports suggest that receipt of systemic therapy after CNS recurrence is an important predictor of survival.7, 10 In our study, of the 62 patients for whom information regarding the use of post-SRS systemic therapy was available, 52 (82%) received further treatment. Breast cancer subtype was assigned based on IHC receptor status acquired from primary breast tumors. Discordance between primary tumor and brain metastasis receptor status has been reported.20 However, lack of availability and safety of CNS biopsies makes this a reasonable surrogate. Finally, one may hypothesize that SRS use decreases the likelihood of neurologic symptoms and possibly neurologic death; however, conclusions regarding this cannot be drawn from this study.
Breast cancer is not well represented in published phase 3 randomized studies of SRS, comprising 6.8% to 11.7% of the study populations.11-13 Several retrospective studies have addressed the use of SRS in breast cancer-specific populations with newly diagnosed brain metastases; however, none exclusively addresses the question of SRS in the salvage setting.14, 21-24 Chao et al previously reported the effective use of salvage SRS after prior WBRT in a population of mixed primary tumor sites.25
This study adds to the considerable available literature reporting that brain metastases from breast cancer are heterogeneous and differ in terms of tumor biology and prognosis. Previous studies have demonstrated that HER2-positive disease is a favorable prognostic factor in patients with newly diagnosed brain metastases.7-11 Our study suggests that HER2 positivity also has prognostic significance at the time of CNS recurrence. This is likely reflective of effective systemic therapy options in the HER2-positive population. Further studies should be performed to define subpopulations of breast cancer patients with brain metastases who are most likely to benefit, but also to define subpopulations who do not benefit from this technique and who may be better served by trials of systemic therapy or by a transition to supportive care.
In patients with good KPS, salvage SRS for breast cancer brain metastases is a reasonable treatment option, given an associated median survival >9 months. Furthermore, patients with HER2-positive tumors at diagnosis or stable extracranial disease at the time of SRS have an improved clinical course, with median survival of >1 year.
This study was supported by the Breast Cancer Research Foundation and the Susan G. Komen Foundation grant BCTR0600849.