• Gamma knife radiosurgery;
  • linear accelerator;
  • surgery;
  • chemotherapy;
  • recursive partitioning analysis;
  • survival


  1. Top of page
  2. Abstract


This study was designed to evaluate the therapeutic effect of stereotactic radiosurgery (SRS) in patients aged ≥75 years who presented with brain metastasis.


The authors analyzed the data from 44 consecutive patients treated with SRS for ≥1 brain metastasis. The median age at the time of treatment for brain metastases was 79.3 years (range, 75 years–86 years), and the median Karnofsky performance status was 80 (range, 50–100). At the time of SRS, 31 patients were treated for a single metastasis, and the remaining 13 patients were treated for ≥2 lesions (n = 74 lesions). The median tumor volume was 1.2 cm3 (range, 0.007 cm3-22.5 cm3). The median maximal and marginal doses were 36 grays (Gy) (range, 18.8 Gy-48.2 Gy) and 20 Gy (range, 10 Gy-24 Gy), respectively.


Median survival was 7.3 ± 1.65 months (range, 1.6 months-38.9 months) from the time of diagnosis of brain metastasis. Median survival of the patients with a single brain metastasis (10.1±1.92 months) was longer than that of the patients with ≥2 metastases (6.6±1.28 months) (P <.02). A single lesion was found to be an independent favorable prognostic factor (P±=±.017; odds ratio, 2.385 [95% confidence interval, 1.167-4.874]) in univariate and multivariate analysis. Patients with nonsmall cell lung cancer fared worse than patients with other tumor types (survival of 6.5 ± 0.70 months vs 10.1 ± 2.33 months [P<.05]).


SRS for patients aged ≥75 years with brain metastases is an effective and safe treatment modality that appears to improve survival, with outcomes that compare favorably with those reported for younger patients in an appropriately selected population. Cancer 2008. © 2008 American Cancer Society.

Cancer can occur at all ages, but its incidence increases with age. In a cross-national comparison study in 7 industrial nations, the proportion of persons aged ≥65 years ranged from 12.6% to 18.1% in 2000.1 It has been estimated that the proportion of elderly persons may reach between 20% and 28% of the population by 2030.2 Furthermore, a person in average health surviving to age 75 years can now expect to live, on average, another 11 years, and a healthy 85-year-old has a 6-year median life expectancy.2, 3

Depending on the type of malignancy, the stage of disease at the time of presentation, and a host of tumor- and treatment-related factors, the incidence of brain metastasis has been reported to range from 5% to 40% of patients with systemic cancer.4-6 There has been a trend toward an increasing incidence of brain metastasis for several reasons, including improved survival in patients with localized cancer due to earlier diagnosis; improved detection of metastatic disease by more sensitive imaging and heightened surveillance; and novel therapeutic strategies, the majority of which do not cross the blood–brain/blood–brain tumor barrier—among others. Although there remains much to be done to understand the biology and epidemiology of brain metastasis of solid epithelial malignancies, several recent, prospective studies have demonstrated the survival advantage of treating brain metastases with surgery and whole brain radiation therapy (WBRT) or WBRT and stereotactic radiosurgery (SRS). However, these prospective studies—and some retrospective reports—suggest the strong, negative influence of age on prognosis, especially in patients aged ≥65 years, and to our knowledge there has been no detailed examination of the utility of SRS, alone or as part of a multidisciplinary approach to treating brain metastases in the elderly (ie, patients aged ≥75 years at the time of diagnosis of brain metastasis).

Combined with the known increase in the incidence of most cancers with age, these advances in life expectancy are likely to intensify the burden of cancer in general and of central nervous system (CNS) involvement by systemic cancer in particular. However, because older adults are more likely to have multiple, concurrent diseases, restricted physiological reserve, limitations in physical functioning, and other age-associated problems that may circumscribe their tolerance to some therapies, treatment options have commonly been regarded as more restricted in patients aged ≥75 years. Thus, in spite of the known efficacy of SRS in treating patients with brain metastases, little is known of its utility in older patients.7, 8 As life expectancies continue to improve and the proportion of patients in this age group continues to rise, this is an important question to address.

Therefore, we performed a retrospective analysis of a consecutive series of patients aged ≥75 years with brain metastases treated with SRS to assess the long-term outcome (minimum follow-up, 2.5 years) and compared our outcomes with previous, prospective studies of younger patients.


  1. Top of page
  2. Abstract

Patient Characteristics

A total of 615 patients with brain metastases underwent SRS at the Cleveland Clinic for the treatment of single or multiple brain metastases between August 1991 and June 2004. Independent, institutional review board–approved databases maintained by the Brain Tumor Neuro-Oncology Center and the Department of Radiation Oncology were queried for patients with metastatic brain tumors who were aged ≥75 years at the time of presentation of brain metastasis and who received SRS at the Cleveland Clinic; 44 (7.2% of the entire population) patients were eligible for inclusion. Patients who were treated with SRS may not represent all eligible patients due to referral patterns within our institution as well as within the region, which serves as a potential selection bias. Data necessary for analyses were extracted, compiled, and verified against patients' archived medical records. The Social Security Death Index was used to verify dates of death when necessary. The Cleveland Clinic Institutional Review Board approved this study.

There were 26 men and 18 women, with a median age of 79.3 ± 3.2 years (range, 75 years-86 years). The median Karnofsky performance status (KPS) score was 80 (range, 50-100). According to the recursive partitioning analysis (RPA) classification of Gaspar et al,8 36 (81.8%) patients were class 2, and 8 were class 3. The primary tumor site was the lung in 20 (nonsmall cell lung cancer [NSCLC] in 17 and small cell lung cancer [SCLC] in 3) patients, genitourinary tract in 9 (renal in 6, prostate in 2, and bladder in 1) patients, gastrointestinal tract in 5 patients, melanoma in 4 patients, breast in 3 patients, and unknown in 3 patients. Excluding 3 patients with an unknown primary cancer site and 4 patients in whom presentation with a brain metastasis led to the identification of the primary, systemic cancer, the median interval between diagnosis of the primary tumor and diagnosis of brain metastases was 22 months (range, 0.4 months-180.1 months; n = 37). At the time of SRS, 25 (56.8%) patients had control of the primary tumor; 19 (43.2%) patients had systemic disease that required ongoing therapy. Nineteen (43.2%) patients had extracranial metastases at the time of SRS, and 25 patients did not have extracranial metastases.

Nineteen of 44 (43.2%) patients received scheduled or salvage external-beam WBRT. Among these 19, upfront WBRT (interval <1 month between WBRT and SRS) was performed in 8 patients, and salvage WBRT was administered to 2 patients. The remaining 9 patients received WBRT first, followed by SRS. The median interval from WBRT to SRS in those 9 patients was 2 months (range, 1.6 months-16.8 months). WBRT doses ranged from 18 to 40 grays (Gy), administered in fractions that ranged from 1.5 to 3 Gy daily. Patients' characteristics and treatment methods are summarized in Table 1.

Table 1. Patient Characteristics and Treatment Methods
CharacteristicNo. of Patients (%)
  1. SRS indicates stereotactic radiosurgery; KPS, Karnofsky performance status; RPA, recursive partitioning analysis; NSCLC, non–small cell lung cancer, WBRT, whole-brain radiation therapy; LINAC, linear accelerator radiosurgery; GKRS, Gamma Knife radiosurgery.

 Male26 (59.1)
 Female18 (40.9)
Median age, y79.3
 ≥75, <8025 (56.8)
 ≥8019 (43.2)
Pre-SRS KPS (median, 80)
 <8018 (40.9)
 ≥8026 (59.1)
RPA class
 II36 (81.8)
 III8 (18.2)
No. of metastases
 Single31 (70.5)
 Multiple13 (29.5)
Primary tumor type
 NSCLC17 (38.6)
 Others27 (61.4)
Primary tumor status
 Controlled25 (56.8)
 Not controlled19 (43.2)
Extracranial metastases
 Absent25 (56.8)
 Present19 (43.2)
Combined WBRT
 Yes19 (43.2)
 No25 (56.8)
Type of SRS
 LINAC9 (20.5)
 GKRS35 (79.5)

Brain Metastases and Radiosurgical Treatment

A total of 74 brain metastases in 44 patients were treated with SRS. For the SRS procedure, all patients except 3 had combined contrast-enhanced magnetic resonance imaging (MRI) and computed tomography (CT) of the brain; the exceptions were 3 patients with pacemakers, in whom MRI was precluded. Contrast-enhanced MRI was performed in all but these 3 patients at least once after radiosurgery. At the time of SRS, 31 patients (70.5%) were treated for a single lesion, and 13 (29.5%) patients had ≥2 lesions (range, 2-9 lesions). The location of brain metastases was supratentorial in 60 patients and infratentorial in 14 patients; none of the latter group of metastases was located within the brainstem. The median volume of lesions treated at SRS was 1.2 cm3 (range, 0.007 cm3-22.5 cm3), and the median longest dimension, in 3-dimensional measurements, was 16.4 mm (range, 3.3 mm-40.9 mm). Until October 1996, linear accelerator (LINAC)-based radiosurgery was performed (n = 9 patients and n = 11 lesions). Subsequently, 35 patients (n = 63 lesions) were treated with gamma knife radiosurgery (GKRS) using the Leksell Gamma Knife, model B or C (Elekta AB, Stockholm, Sweden). The median maximal dose was 36 Gy (range, 18.8 Gy-48.2 Gy), with a marginal tumor dose ranging from 10 Gy to 24 Gy (median, 20 Gy) at the 50% to 90% isodose line. For patients undergoing SRS for ≥2 lesions (in this series, up to 9 lesions at a single setting; no multiple treatment settings were needed), standard multiple matrix planning algorithms were used to insure that no lesion was overtreated and to minimize exposure to the remainder of the CNS, according to standard SRS planning principles.

Survival Time and Statistical Analysis

Survival was calculated from the date of the first neuroimaging study (MRI or CT scan) that documented brain metastases and led to treatment until death. Statistical analyses were performed using the SPSS program (version 10.0 for Windows; SPSS Inc, Chicago, Ill). The Kaplan-Meier method was used to predict overall survival. For univariate analysis, the log-rank test was used for categoric variables, and the Cox proportional hazards model was used for continuous variables. Continuous variables are presented as mean ± standard error. Factors that were evaluated for their potential influence on survival included age, sex, KPS, RPA classification, number of brain metastases (single vs multiple), primary tumor sites, primary tumor status, presence of extracranial metastases, and combined WBRT. A final multivariate analysis was calculated using a stepwise forward logistic regression. A P value <.05 was considered statistically significant.


  1. Top of page
  2. Abstract


The mean follow-up duration from the time of initial diagnosis of a brain metastasis was 10.1 ± 7.3 months (range, 1.6 months-62.9 months). During the follow-up period, 43 of 44 patients died, and 1 patient was alive at the time of last follow-up, 16.7 months after the diagnosis of brain metastasis. Overall median survival from the diagnosis of brain metastases was 7.3 ± 1.7 months (range, 1.6 months-62.9 months) (Fig. 1). Actuarial overall survival rates at 3 months, 6 months, 9 months, and 12 months were 86.4 ± 5.2 %, 68.2 ± 7.0 %, 45.5 ± 7.5 %, and 34.1 ± 7.2 %, respectively.

thumbnail image

Figure 1. Overall survival is shown from the diagnosis of brain metastases in 44 patients aged ≥75 years at the time of treatment with stereotactic radiosurgery. Overall median survival from the diagnosis of brain metastases was 7.3 months. Actuarial overall survival rates at 3 months, 6 months, 9 months, and 12 months were 86.4 ± 5.2%, 68.2 ± 7.0%, 45.5 ± 7.5%, and 34.1 ± 7.2%, respectively.

Download figure to PowerPoint

Factors Correlated With Survival Time

The results of analyses for variables that could be correlated with survival time are shown in Table 2. In univariate analysis, the number of brain metastases (single vs multiple) at the time of SRS and the primary tumor type (NSCLC vs others) influenced survival significantly. The median survival time of 31 patients with a single brain metastasis (10.1 ± 1.92 months) was significantly longer than that of 13 patients with ≥2 brain metastases (6.6 ± 1.28 months) (Fig. 2); survival in patients with only 2 metastases was not found to be better than in those with ≥3 metastases (data not shown). Among 31 patients with a single metastasis, 2 patients survived >2 years after diagnosis of the brain metastasis. One patient (aged 75 years) with renal cell cancer treated with SRS alone survived for 26.5 months; the other patient (aged 79 years), who had SCLC treated with WBRT, followed by SRS 3 months later for residual disease, survived 38.9 months.

thumbnail image

Figure 2. Survival is shown according to the number of brain metastases (single vs multiple). Patients with a single metastasis fared better than patients with ≥2 lesions. Actuarial median survival for patients with a single versus multiple metastases was 10.1 months ± 1.92 months versus 6.6 months ± 1.28 months (P = .017 by multivariate, forward stepwise analysis; odds ratio, 2.385 [95% confidence interval, 1.167-∼4.874]).

Download figure to PowerPoint

Table 2. Analysis of Prognostic Factors Related to Survival
Prognostic FactorP Value
Univariate Analysis*Multivariate Analysis
  • NSCLC indicates nonsmall cell lung cancer; NS, not significant; KPS, Karnofsky performance status; RPA, recursive partitioning analysis; WBRT, whole brain radiation therapy.

  • *

    Log-rank test for categoric variables and Cox proportional hazards model for continuous variables (age).

  • Cox proportional hazards model (forward stepwise method).

  • Significant factors (P<.05) in both univariate and multivariate analysis are identified in bold for ease of identification.

No. of brain metastases (single vs multiple).014.017
Primary tumor type (NSCLC vs others).040NS
Primary tumor status (controlled vs not).204NS
Extracranial metastases (absent vs present).645NS
Sex (male vs female).784NS
KPS (≥80 vs <80).649NS
RPA class (II vs III).475NS
Combined WBRT (yes vs no).863NS
Age, y (75-79 vs ≥80).776NS

The type of primary tumor site, when classified by anatomic location (lung, gastrointestinal, genitourinary, etc) did not appear to significantly affect patients' survival time. However, the median survival time of 17 patients whose original cancer was NSCLC (6.5 months ± 0.70 months) was shorter than that of the other 27 patients (10.1 months ± 2.33 months) (P < .015) (Table 2) (Fig. 3). Eleven of 17 (64.7%) patients with NSCLC had a single lesion at the time of SRS, whereas 20 (74.1%) of 27 patients with other primary tumors, including those with an unknown primary tumor, had a single lesion. In multivariate analysis, only the number of brain metastases (single vs multiple) was found to influence survival significantly (P = .017; odds ratio, 2.385 [95% confidence interval, 1.167-∼4.874, using the forward stepwise method]). Other factors, such as age, sex, KPS, RPA classification, primary tumor status, presence of extracranial metastases, and combined WBRT, did not appear to influence survival significantly (Table 2). The median KPS at an average of 3 months after SRS was 80 (range, 60-100); no patient had a permanent decrease in KPS after SRS.

thumbnail image

Figure 3. Survival is shown according to the primary tumor type (nonsmall cell lung cancer [NSCLC] vs others). Patients with NSCLC fared less well than patients with any other histologic subtype, alone or in combination. Actuarial median survival for NSCLC versus others was 6.5 months ± 0.70 months versus 10.1 months ± 2.33 months (P = .0401 by univariate log-rank test) (see Table 2).

Download figure to PowerPoint

Causes of Death

As noted, 43 (97.7 %) of 44 patients had died at the time of last follow-up. In 28 (65.1%) cases in which a cause could be determined, 23 (82.1%) patients died of systemic disease progression, 4 (14.2%) died of CNS progression (2 with new lesions and 1 at the local site), and 1 (3.6%) died of a myocardial infarction. In 15 cases, the cause of death was unknown or could not be determined precisely.

Treatment-Related Complications

Among the 44 patients with 74 brain metastases, at least 1 neuroradiologic imaging examination in 26 (59.1%) patients was performed at the Cleveland Clinic and was reviewed for this study. At a mean follow-up of 3.8 months ± 5.8 months, 5 (19.2%) of 26 patients developed a new, distant brain metastasis, which was treated successfully with SRS in 4 patients who returned to Cleveland Clinic for follow-up. During this same period, 5 (12.2%) of 41 patients treated for metastases demonstrated local recurrence after SRS treatment, and were subsequently treated with WBRT, with control obtained in 3 patients for whom follow-up imaging and information was available; in 1 patient, the lesion progressed and caused the patient's death. The median interval from SRS to distant failure was 4.8 months ± 22.1 months (range 2.5 months-53.9 months), and the median interval from SRS to local failure was 5.4 months ± 8.1 months (range, 2.9 months-22.5 months).

One patient experienced apparent radiation necrosis. An 83-year-old man was diagnosed with esophageal cancer 16.7 months before the development of a single brain lesion with a volume of 10.9 cm3 and a maximal dimension of 3.29 cm. He underwent GKRS only and received 15 Gy to the 50% isodose line. Six months after GKRS, imbalance and gait disturbances developed; radiation necrosis was suspected on MRI. A short course of corticosteroids resulted in clinical improvement; follow-up imaging 6 months later demonstrated near complete resolution of the enhancing lesion, and he was medication free. The patient died 18 months after SRS of systemic disease progression.


  1. Top of page
  2. Abstract

Age has been shown to be an important prognostic factor in the majority of patients with malignant primary or metastatic brain tumors. Many studies suggest that elderly patients have a worse outcome compared with younger patients.8-12 Although a precise definition of elderly has not been determined, many researchers have chosen 65 years because it is the cutoff selected for RPA classification.8

To our knowledge, to date only a few studies have been published that have attempted to assess outcomes or survival in elderly patients with brain metastases. In 1 retrospective series, 916 patients with brain metastases were treated with postoperative or definite WBRT over a period of 21 years. The median overall survival was 3.8 months from the first day of WBRT for patients aged <65 years and 2.6 months for patients aged ≥65 years (P = .0002). In a second study, Noel et al7 retrospectively analyzed the outcomes of LINAC SRS for brain metastases in 117 patients aged ≥65 years; overall median survival time was 8 months from the patient's first SRS. In the series, the median survival of elderly patients aged ≥75 years who were treated with SRS was 7.3 months (range, 1.6 months-38.9 months) from the diagnosis of brain metastases, which is comparable to the somewhat younger population in the study by Noel et al. By comparison, in the Radiation Therapy Oncology Group (RTOG) 9508 randomized trial of SRS with or without WBRT, the median survival of patients with a single brain metastasis, treated with both modalities, was 6.5 months.13 Although the current series had fewer cases and a narrow age range (and we were unable to control for the possibility that there may have been patients not sent to us, due to referral bias, whose outcomes could have been different from what we observed in this cohort of patients), there was also no significant association found between age and survival time in our elderly group. Finally, in the vast majority of cases (85.1%), the cause of death was either systemic disease progression (81.4%) or not related to the tumor (3.7%). These findings suggest that patients aged ≥75 years are likely to benefit equally when compared with younger patients from aggressive treatment strategies for their CNS metastases.

In addition to age, performance status, extracranial tumor activity, control of primary tumor, and number of brain metastases have been shown to be prognostic factors in patients with brain metastases.5, 7, 8, 10, 11, 13-15 Among those factors, there is a consensus regarding the survival benefit with single brain metastases compared with multiple lesions. The University of Minnesota published a retrospective review of 729 patients with brain metastases treated over a 20-year period.5 They reported that median survival from presentation with brain metastases was 5 months for 384 (53%) patients with single lesion and 3 months for 345 (47%) patients with multifocal disease. Similar general findings have been noted in prospective trials of WBRT, alone or with SRS,8, 13 in retrospective trials,5, 7, 10, 11 and in retrospective surgical series.16 In the current series, only the number of brain metastases (single vs multiple) was found to be an independent prognostic factor of survival.

WBRT has been a standard treatment for brain metastases for several decades.4, 6, 17 However, multiple logistic barriers may prevent elderly patients from receiving radiation therapy, including limited access to transportation, impaired functional or cognitive status, and poor social support.18, 19 Although the prophylactic role of WBRT has already been demonstrated in some cancers,20 the issue of the duration of treatment, with multiple fractions, may be a reasonable consideration for elderly patients. Increasingly, the efficacy and safety of SRS, alone or combined with WBRT, has been recognized.8, 13, 21 A multi-institutional retrospective study of patients initially treated with SRS alone (24% of whom ultimately received salvage WBRT) versus SRS with upfront WBRT (interval between WBRT and SRS of ≤1 month) in 569 patients demonstrated that there was no survival difference between the 2 groups when adjusted for RPA class.21 Another multi-institutional, prospective, randomized controlled trial compared WBRT plus SRS versus SRS alone for patients with 1 to 4 brain metastases and also indicated no significant difference in median survival time (7.5 months vs 8.0 months) between the 2 groups.22 In this study, SRS alone without upfront WBRT was associated with increased brain tumor recurrence, which necessitated salvage treatment more frequently, but salvage WBRT did not result in either a worsened neurologic outcome or an increased risk of neurologic death.22 Finally, the RTOG 9508 study demonstrated that a radiosurgery boost after WBRT improves survival compared with WBRT alone for surgically unresectable single brain metastases (median survival time of 6.5 months vs 4.9 months; P = .0393 in univariate analysis).13 This study also demonstated a statistically significant improvement in KPS and decreased steroid use at 6 months in the SRS boost treatment group. In this trial, they noted that radiosurgery boost after WBRT did not increase treatment-related acute toxicities compared with WBRT alone in the subset of patients aged ≥75 years (23 of 160 patients [20%] vs 37 of 166 patients [22%]). In the current series, symptomatic radiation necrosis was observed in only 1 patient who underwent SRS only, and it was controlled with a short course of corticosteroids, with resolution. None of our 19 patients who received both SRS and WBRT developed radiation necrosis. Although to our knowledge there is no general consensus regarding the risks and benefits of omitting upfront WBRT, it appears that salvage WBRT or repeated radiosurgery may be effective as salvage therapy for recurrent tumors after SRS alone, especially in elderly patients.

Although there are limitations to this study, such as selection bias in treatment (including SRS), varied approaches to treatment of the CNS lesion(s), uncontrolled mix of histologic subtypes, and lack of control of the medical oncologists' treatment of the systemic disease, it does reflect the contemporary situation of patients with brain metastasis in the community. Any patient with systemic disease who is stable or receiving therapy, for whom the oncologist predicts a life expectancy of ≥3 months, is considered for treatment, which may include surgery, SRS, and WBRT. The results of the current study suggest that the majority of patients aged ≥75 years have outcomes equal to those of younger patients, with good performance status, and no increased risk of adverse effects.

Finally, although the expected survival of patients with brain metastases may be limited in the majority of cases in the long term, selected subsets of the patients can have prolonged and good-quality survival. Kondziolka et al compared 44 patients who survived for >4 years after radiosurgery with a cohort of 100 patients who lived for <3 months after radiosurgery.15 In their study, long-term survivors (44 of a total of 677 patients; 6.5%) had a higher initial KPS, fewer brain metastases, and less extracranial disease at the time of presentation.15 Recently, Chao et al23 reported that 32 of 1288 (2.5%) patients survived ≥5 years, with a median survival of 9.3 years, after diagnosis of their brain metastases.23 They concluded that for patients with good prognostic factors such as young age (aged <65 years), good RPA characteristics (class 1), and a single metastasis, aggressive treatment with surgery or SRS may offer the best chance for long-term survival. In the current series, 2 out of 44 (4.5%) elderly patients survived >2 years after the diagnosis of a single brain metastasis, with both patients functioning independently.

In summary, the current study results suggest that stereotactic radiosurgical treatment of patients aged ≥75 years with brain metastases, with a good functional status, less extra-CNS disease, and a reasonable expectation for survival after SRS (for example, in this cohort of patients, at least 3 months), is an effective and safe treatment modality that prolongs survival, has an excellent functional outcome, and provides results that are similar to those noted in younger patients. Based on our outcomes, we will continue to offer SRS for patients with brain metastasis who are aged ≥75 years.


  1. Top of page
  2. Abstract