• breast carcinoma;
  • brain metastases;
  • radiosurgery;
  • gamma knife;
  • radiotherapy


  1. Top of page
  2. Abstract


The current study analyzed the feasibility and outcome of stereotactic radiosurgery (SRS) for treatment of brain metastases from breast carcinoma.


During an 8-year period, 151 patients with a combined total of 620 brain metastases from breast carcinoma underwent 197 outpatient SRS procedures. Sixty-three percent of all patients had multiple brain metastases. The median tumor volume was 2.2 cm3 (range, 0.1–20.9 cm3). The mean prescribed tumor dose was 19 ± 4 grays. Local/distant tumor recurrences were treated with additional radiosurgical therapy for patients with stable systemic disease. All patients were categorized according to the Radiation Therapy Oncology Group classification. Survival time and freedom from local tumor recurrence were analyzed using the Kaplan–Meier method. Prognostic factors were identified using the Cox proportional hazards model.


The overall median survival duration was 10 months after SRS. Ninety-four percent of patients did not experience local brain tumor recurrence after radiosurgery. In addition, 70.2% of patients did not have disease recurrence in the brain. Most patients died of systemically progressing malignancy. A Karnofsky performance score > 70 and recursive partitioning analysis Class I were related to prolonged survival in the univariate and multivariate analyses. Age, whole-brain radiotherapy, surgery, number of metastases, chemotherapy, and latency period from diagnosis of the primary tumor to the development of brain metastases did not reach prognostic relevance in the multivariate model. Patients with RPA I, II, and III survived 34.9, 9.1, and 7.9 months, respectively. There was no treatment related permanent morbidity and mortality. The transient morbidity rate was 17%. Sixteen patients exhibited symptomatic transient complications related to treatment.


The results of the current study indicate that SRS is a feasible treatment concept for selected patients with multiple brain metastases from breast carcinoma. Cancer 2004. © 2004 American Cancer Society.

Brain metastases in patients with breast carcinoma indicate a terminal stage of the disease. Brain metastases occur clinically in 10–15% and at autopsy in 20–30% of patients with breast carcinoma.1, 2 Historically, breast carcinoma has been considered to be a relatively radiosensitive tumor. Therefore, whole-brain radiotherapy (WBRT) is the cornerstone of treatment for patients with multiple and/or nonsurgically accessible brain metastases.1, 3 However, the overall prognosis after treatment of patients with brain metastases from breast carcinoma remains unclear, due to vaguely defined selection criteria (e.g., number of tumors, tumor size), insufficient follow-up data, and interactions among applied treatment regimens such as WBRT, surgery, and/or stereotactic radiosurgery (SRS) in most retrospective studies.3–6 The current retrospective, single-institution analysis was conducted to present for the first time detailed outcome data (adjusted for the effects of prognostic factors) for a large, well defined patient population (151 patients with a combined total of 620 tumors) with brain metastases from breast carcinoma undergoing SRS over a study period of 8 years. Our goal was to determine whether a local outpatient treatment (SRS) can control multiple brain metastases in selected patients with breast carcinoma.


  1. Top of page
  2. Abstract

Between September 1994 and January 2003, 151 patients with 620 cerebral metastases from breast carcinoma underwent 197 gamma knife procedures (Leksell Gamma Knife; Elekta Inc., Atlanta, GA). One hundred seventy-one patients with brain metastases from breast carcinoma were selected for radiosurgery according to the following eligibility criteria: 1) diagnosis of breast carcinoma; 2) histologic verification of brain metastases by stereotactic biopsy for patients with uncertain diagnoses; 3) maximum tumor diameter ≤ 3 cm; 4) probable life expectancy ≥ 3 months; 5) Karnofsky performance score (KPS) ≥ 50; 6) extracranial tumor that was stable or in remission with or without systemic therapy; and 7) exclusion of meningeal or ependymal tumor spread by magnetic resonance imaging (MRI) and/or cerebrospinal fluid examination.

Gamma knife surgery (GKS) was performed consecutively by a specialized neurosurgeon as an outpatient procedure on the basis of stereotactic MRI scans for all patients. Patient data were collected prospectively in a computerized data base. Sixty-three percent of all patients had multiple cerebral metastases. Thirty patients underwent surgery before GKS to resect large metastases not eligible for radiosurgery. Patients who received WBRT before SRS were not excluded form the study. The histology of the primary tumor was an infiltrating ductal or lobular carcinoma in most patients. Gadolinium-enhanced MRI scans were used for treatment planning as well as for the follow-up examinations of all patients. Tables 1 and 2 provide detailed summaries of patient and treatment characteristics, respectively. Multiple isocenters were chosen to match the tumor volume as accurately as possible. For local or distant tumor recurrences, an additional radiosurgical procedure was performed if the patient was clinically stable. SRS treatment parameters are shown in Table 3. Side effects were characterized as asymptomatic if they were demonstrated on MRI scans and if patients did not show clinical symptoms. Symptomatic or lethal effects were scored similarly. Side effects were considered due to either radiation toxicity (RT) to brain tissue or hemorrhages from treated metastases.

Table 1. Patient Characteristics
CharacteristicNo. of patients/events (%)Median (range)
  1. KPS: Karnofsky performance score; GKS: gamma knife surgery; RPA: recursive partitioning analysis.

All patients151 
 Female148 (98) 
 Male  3 (2) 
Age (yrs) 60 (35–78)
KPS 80 (40–100)
Neurologic deficits 79 (55) 
Brain metastases  
 Total no.620 
 No. per patient  3 (1–20)
 Single 56 (37) 
 Multiple 95 (63) 
 No. per GKS  2 (1–20)
RPA class  
 I 18 (12) 
 II109 (72) 
 III 24 (16) 
Extracranial metastases133 (88) 
Histologic verification of metastases to the brain  
 By stereotactic biopsy 11 (7) 
 By brain tumor resection 30 (20) 
Volume of metastases (cm3) 2.2 (0.1–20.9)
Table 2. Treatment Characteristics
CharacteristicNo. of patients
  • GKS: gamma knife surgery; WBRT: whole-brain radiotherapy.

  • a

    Mean ± standard deviation.

No. of GKS sessions per patient197
Time to brain metastases (yrs)6.1 ± 0.8a
Table 3. Radiosurgery Treatment Parameters
ParameterMean ± SDMedian (range)
  1. Gy: grays; SD: standard deviation.

Dose (Gy)  
 Maximum37 ± 436 (26–47)
 Minimum19 ± 418 (10–36)
Peripheral isodose (%)52 ± 650 (40–95)
No. of isocenters/metastases4.8 ± 3.1 4 (1–13)

Follow-Up Evaluation

Neurologic examination and tumor response as verified by stereotactic MRI scans were used to evaluate patients during follow-up. Follow-up examinations were performed at 3 and 6 months after radiosurgery and then every 6–9 months until death or the date of closure of the study (January 31, 2003). KPS was used to define improvement/deterioration; a KPS that was better/worse relative to preoperative findings indicated an improved/deteriorated status. Otherwise, the patient's status was considered to be stabilized. Local disease recurrence was defined as the reappearance of a metastasis at exactly the same site as the first metastasis, and distant disease recurrence was defined as the appearance of a new brain metastasis at a site different from that of the original metastasis detected by the MRI scan.

Tumor volume on follow-up MRI scans was calculated using a computerized planning system. The cause of death was determined from medical records and from the referring physician's correspondence or supplementary phone calls. The cause of death was determined using the prospective study protocol of Patchell et al.7 That is, patients with stable extracerebral disease and progressive neurologic dysfunction, patients with severe neurologic disability dying from intercurrent illness, and patients with progressive systemic and neurologic disease were considered to have experienced neurologic death. Otherwise, a systemic death was assumed.7 Autopsy data were not available.

Statistical Methods

The reference point of the study was the date of the radiosurgical procedure. The endpoints were death and date of local disease recurrence. The length of survival and freedom from local disease recurrence were estimated using the Kaplan–Meier method.8 Comparison of Kaplan–Meier curves was performed with the log-rank statistic. The prognostic value of the individual covariates was obtained from the Cox proportional hazards model.9 Variables used for univariate and multivariate analyses were dichotomized. The correlation between prognostic factors was analyzed using the chi-square statistic. First, in the prognostic model, the importance of each covariate was tested by univariate analysis. Second, all variables were fitted together (full model). The best model contained only variables associated with the length of survival. The following variables were tested: age at radiosurgery (> 65 years vs. ≤ 65 years); pretreatment KPS (≤ 70 vs. > 70); number of brain metastases (1 vs. > 1); WBRT (yes vs. no); surgery (yes vs. no); immunotherapy (yes vs. no); recursive partitioning analysis (RPA) class (I vs. II vs. III); and latency period between diagnosis of the primary tumor and development of cerebral metastasis (< 1 year vs. ≥ 1 year).


  1. Top of page
  2. Abstract

Follow-up information was available for all patients. The median follow-up period was 8.3 months (range, 1 week to 55 months). The status of the primary tumor was classified as disseminated (RPA Class II/III) in 133 patients. Thirty-seven patients (24.5%) were transferred to our institution after WBRT (30–40 Gy) was performed. Nine patients received WBRT after SRS (5.9%). The diagnosis of brain metastasis was based on radiologic findings and the diagnosis of the primary tumor. Eleven patients underwent stereotactic biopsy before SRS, and 30 patients underwent surgical resection for tumors > 3 cm in diameter. Seventy-nine patients presented with neurologic symptoms, such as seizures or focal neurologic deficits, before SRS. All tumor locations in the brain were treated. In 21 patients, the tumor was located in the midbrain/brainstem. Extracranial metastases were surgically removed in 27 patients. Ninety patients received radiotherapy directed to areas outside the brain.

Prognostic Factors for Survival

One hundred ten patients had died at the time of the last follow-up. Nine patients died of progressive central nervous system disease (new distant metastases). For 21 patients who developed active disease outside the brain and some form of neurologic deterioration, it was difficult to determine whether death was attributable to intracranial or extracranial disease. These patients were considered to have experienced neurologic death. Aside from eight patients who had unknown causes of death, all patients died of progressive systemic disease. The median survival time (MST) was 10 months (Fig. 1). Patients with controlled systemic disease and no extracranial metastases (RPA Class I) had an MST of 34.9 months (Fig. 2). The maximum (minimum) length of survival was 55 months (2 days). KPS > 70 and RPA Class I were prognostically significant on univariate and multivariate analyses. Advanced age (≥ 65 years) was a negative predictor of survival in the univariate analysis but lost its significance in the multivariate model. All other prognostic factors, including the number of treated metastases (Fig. 3) and WBRT, did not achieve prognostic relevance (Table 4).

thumbnail image

Figure 1. Graph showing cumulative survival rates after stereotactic radiosurgery for all 151 patients in the current study.

Download figure to PowerPoint

thumbnail image

Figure 2. Kaplan–Meier survival by recursive partitioning analysis (RPA) class. Survival differences were statistically significant in the univariate and multivariate analyses (P = 0.0012). The median survival times for patients in RPA Classes I, II, and III were 34.9, 9.1, and 7.9 months, respectively.

Download figure to PowerPoint

thumbnail image

Figure 3. Kaplan–Meier survival by number of brain metastases. The difference between groups was not statistically significant (P = 0.9).

Download figure to PowerPoint

Table 4. Prognostic Factors
VariableNo. of patientsMST (mos)aLog-rank P valueCox model P value (exp [coef])
  • MST: median survival time; KPS: Karnofsky performance score; WBRT: whole-brain radiotherapy; RPA: recursive partitioning analysis exp: exponent; coef: coefficient.

  • a

    ± standard deviation.

Age (yrs)  < 0.0410.17 (1.4)
 ≤ 653711.0 ± 3.0  
 > 651147.9 ± 1.7  
KPS  0.0030.04 (0.5)
 ≤ 70257.8 ± 2.4  
 > 7012610.5 ± 1.4  
No. of metastases  0.90.36 (0.8)
 1569.5 ± 1.9  
 > 19510.0 ± 1.6  
WBRT (1.0)
 Yes4611.4 ± 3.5  
 No1059.5 ± 1.4  
Surgery  0.10.13 (0.6)
 Yes3014.9 ± 4.5  
 No1218.7 ± 1.2  
Immunotherapy/chemotherapy  0.80.1 (1.5)
 Yes11310.3 ± 1.7  
 No387.9 ± 3.8  
RPA class  < 0.00010.0012 (0.1)
 I1834.9 ± 5.3  
 II1099.1 ± 1.5  
 III247.9 ± 2.4  
Latency period for brain metastases/primary tumor (yrs)  0.90.3 (0.6)
 ≤ 11211.5 ± 4.9  
 > 11399.6 ± 1.2  

Treatment Response

The local brain tumor control rate was 94%. Local brain tumor recurrences were observed in nine patients. The rate of freedom from recurrence in the brain was 70.2%. Radiosurgical retreatment was performed in 29 patients for new, distant metastases and in 5 patients for local and distant tumor recurrences. Up to five repeated SRS treatments were performed. Seven of 37 patients who received WBRT before radiosurgery developed recurrent tumors. In seven patients, an additional radiosurgical treatment was not performed due to a miliar cerebral tumor seeding. Nine patients developed leptomeningeal disease.

Postoperative Course

Of 79 patients who presented with neurologic symptoms before radiosurgery, 55 (70%) had an improved KPS after SRS, 15 (19%) had stable disease, and 9 (11%) had progressive disease (4 patients due to progressive paresis and 5 due to increased seizure frequency) at the first follow-up examination. Symptom improvement or progression occurred within a few days to a few weeks. All patients received steroids for 5–7 days (dexamethasone: 3 × 4 mg). Dose reductions were made based on neurologic status.

Cause of Death Analysis—Complications after Radiosurgery

The overall morbidity and mortality rates were 17% and 0%, respectively (Table 5). Symptomatic RT was observed in 16 patients. RT developed 4 weeks to 20 months after radiosurgery and was associated with transient focal neurologic deficits and/or seizures, which were treated with steroids and/or anticonvulsants. No patient underwent decompressive surgery caused by space-occupying radionecrotic lesions. Asymptomatic RT (typical imaging changes without clinical signs) was observed in nine patients. One patient developed an asymptomatic intratumoral hemorrhage after SRS.

Table 5. Complications and Causes of Death after Stereotactic Radiosurgery for Brain Metastases from Breast Carcinoma
FactorNo. of patients (%)
Treatment-related complications 26 (17)
Hemorrhage  1 (1)
 Lethal  0 (0)
 Symptomatic  0 (0)
 Asymptomatic  1 (1)
Radiation toxicity 25 (17)
 Lethal  0
 Symptomatic 16 (11)
 Asymptomatic  9 (9)
Cause of death 
 Systemic 72 (65)
 Neurologic 30 (27)
 Unknown  8 (7)
 Total110 (73)


  1. Top of page
  2. Abstract

The presence of multiple brain metastases from breast carcinoma portends rapid clinical deterioration. Most patients with these metastases undergo palliative WBRT. However, unsatisfactory overall survival results and relatively high brain failure rates have been reported by various investigators.1, 3, 4 Recently, stereotactically guided high-precision irradiation as a single-dose application (radiosurgery) showed promising treatment results for selected patients with brain metastases.5, 10, 11–17 Radiosurgery is attractive due to its low risk and minimal invasiveness. It can be used in conjunction with, or as an alternative to, other treatment methods and can be performed on an outpatient basis.18 Multiple lesions can be treated simultaneously, and retreatment can be performed for local or distant tumor recurrences. Due to a limited number of reported treatment results after SRS for brain metastases of patients with breast carcinoma, the therapeutic impact on this patient subpopulation remains unclear.5, 10 To our knowledge, the current study represents the first report on a large cohort of patients with metastases from breast carcinoma who received SRS exclusively. Our objective was to assess the specific therapeutic impact of a local treatment concept (SRS) on patients with brain metastases from breast carcinoma.

Treatment Efficacy

A high percentage of local tumor control (94%) was achieved after radiosurgery in the current study. Only 9 of 151 patients did not respond to local treatment. Radiosurgical retreatment was performed for all patients with local or new distant metastases to the brain and a stable systemic tumor status. The overall MST was 10 months. Overall median survival rates ranging from 4 to 16 months have been reported by other authors after WBRT, SRS with/without WBRT, and surgery with WBRT for patients with brain metastases from breast carcinoma.3, 5, 6, 19, 20 These widely varying estimates of survival should be regarded cautiously, as they are influenced by selection bias related to the individually applied treatment regimens. For example, an overall survival of 4–5 months was reported after conventional WBRT.3, 4 However, WBRT groups contained many patients with poor prognosis, including those with unfavorable characteristics (e.g., larger tumors or active systemic disease). Longer survival durations have been reported following surgery and WBRT.19, 20 For example, Pieper et al.19 found an MST of 16 months after surgery and WBRT, which is the highest reported MST following treatment of brain metastases from breast carcinoma; it is clear that a highly selected patient population was involved in that study. The majority of the patients were in good clinical grade and had singular surgically accessible metastases in 87% of the cases.19

One could argue that the true impact of SRS is overshadowed by the effects of systemic chemotherapy (which has been reported to yield favorable responses in selected patients with breast carcinoma) applied to most patients in the current study. It is noteworthy that 90% of patients in the current study received systemic chemotherapy before SRS, which would suggest a relatively low response rate associated with current systemic chemotherapy regimens for brain metastases from breast carcinoma. However, the interaction of systemic chemotherapy and radiosurgery was not further analyzed in the current study.

Prognostic Factors

Data heterogeneity, relatively short follow-up times (resulting in a limited number of events), incomplete evaluation, ill-defined selection criteria for applied treatment strategies, and the use of different (uncomparable) statistical techniques are responsible for the lack of a consensus regarding the relative importance of prognostic factors in patients with brain metastases.22, 27 Mahmoud-Ahmed et al.,3 for example, reported overall treatment results after WBRT for patients with brain metastases from breast carcinoma. Follow-up films were available for 50% of the 116 patients who were selected retrospectively. The current analysis refers to a selected subpopulation with small and primarily multiple brain metastases from breast carcinoma. Most patients in the current study were treated with SRS alone, without additional WBRT. Radiosurgical treatment planning and execution were comparable for all patients. SRS was planned and exclusively performed for all patients by a single neurosurgery specialist. The remarkable finding was that the presence of multiple metastases had no prognostic impact for this selected patient population. Patients with multiple tumors (63%) had their disease controlled with the same efficacy as did patients with single tumors over 1 treatment cycle. This finding does not support the results of several recent surgical, radiosurgical, and radiotherapeutic trials.5, 19, 23 A greater number of metastases was correlated with an inferior prognosis in those studies. It is noteworthy that the WBRT, which is regarded as the standard treatment for patients with multiple brain metastases, did not have a prognostic impact on 46 patients who received additional WBRT. There was no difference in local tumor control or survival time. Several retrospective studies have been published comparing patients treated with SRS who received WBRT with those who did not receive WBRT.24–27 Those studies have generally shown that the disease recurrence rate is reduced by adjuvant WBRT but that the length of survival is not increased.28 Because of the inherent risk of patient selection bias in retrospective studies, this type of study cannot resolve whether patients should receive a combination of WBRT and SRS. Taking into account the results of the current study and the fact that most investigators have suggested a combination of SRS and WBRT for selected patients with brain metastases, we believe that the initiation of a Phase III trial comparing SRS/WBRT with SRS alone is justified. Apart from local tumor control and MST, a Phase III trial should analyze other important endpoints such as quality of life, cost-effectiveness, and neurocognitive sequelae after SRS and WBRT.


Eleven patients developed symptomatic treatment-related complications between 1 and 20 months after SRS; these complications were controlled with steroids. No patient died of radiation-induced complications. Most notably, patients undergoing multiple radiosurgical procedures for local or distant retreatment did not have a higher risk of developing radiogenic complications compared with patients undergoing a single session. In addition, no patient underwent decompressive surgery for space-occupying radionecrotic lesions.


In summary, outpatient stereotactic radiosurgery is a feasible treatment method for controlling small brain metastases from breast carcinoma. Outcome is particularly favorable in patients with a controlled primary tumor and no extracranial disease (RPA Class I). The number of brain metastases has no prognostic relevance. The results of the current study suggest that a Phase III trial to compare SRS/WBRT with SRS alone and examine additional important endpoints, such as quality of life, cost-effectiveness, and neurocognitive sequelae, is justified.


  1. Top of page
  2. Abstract