Efficacy and safety of CyberKnife radiosurgery in elderly patients with brain metastases: a retrospective clinical evaluation

Stereotactic radiosurgery (SRS), in place of whole brain radiation therapy (WBRT), is the standard for patients with 1–3 metastases and an effective treatment for patients with up to 10 lesions to reduce neurotoxicity and preserve quality of life [1‐5]. Brain metastases incidence has been increasing and will continue to rise due to the improvement of systemic therapies [1, 2, 6]. In the current era of immunotherapy or other targeted therapies, overall survival (OS) has increased; thus, the demand for less toxic alternatives than WBRT and the need for retreatment with radiosurgery for local progression or for new cerebral lesions is on the rise [2, 7, 8]. It has been postulated that in the near future, 70% of new cancer diagnoses per year will be reported among the elderly [9, 10]. Consequently, clinicians are facing a larger population of aging patients.

The management of the elderly remains an issue of debate, as they comprise an inhomogeneous patient population with diverse comorbidities and different levels of fitness. Age and the Karnofsky performance score (KPS) have been used to roughly categorize the patients, as proposed by the recursive partitioning analysis (RPA) for prognostic factors by the Radiation Therapy Oncology Group, where patients ≥ 65 years old were classified as patients at medium risk with RPA class II [11]. For patients ≥ 70 years who received GammaKnife SRS of their brain metastases, Park et al. identified the graded prognostic assessment (GPA) classification as a strong prognostic factor for survival [12]. Nevertheless, elderly patients are generally underrepresented in randomized clinical trials, and thus clinical practice is based on small patient series and personal experience and assessment of the treating physician.

Importantly, SRS may represent the treatment of choice considering the risk of cognitive decline in the elderly after WBRT [13, 14]. However, few reports have assessed the feasibility and efficacy of SRS in the elderly using conventional linear accelerator (LINAC) or GammaKnife-based SRS, and to our knowledge, none have done so using a frameless technique with the CyberKnife (CK, Accuray Inc., Sunnyvale, CA), an image-guided robotic linear accelerator system [15‐18]. Thus, our aim in this study was to characterize our elderly patient cohort with brain metastases and to analyze the efficacy and safety of CK-SRS in this population.

Our study evaluated for the first time the efficacy and safety of image-guided frameless CK-SRS in a large cohort of elderly patients. Our study supports the recommendation of SRS for this population in light of the reasonable local control rates with very rare complications.

SRS provided advantages concerning the risk of neurocognitive decline in an elderly population after radiation therapy [13, 14]. Besides, the convenience of a short treatment in an outpatient setting should be considered as an advantage. However, the accumulated experience with SRS for this specific patient population is based on few reports using LINAC or GammaKnife-based SRS [15‐18]. As the OS after cancer diagnosis is on the rise due to new therapeutic regimens, the role of radiosurgery will become more important for the treatment of the elderly [2, 7‐10].

In our series, the median follow-up time was 7.6 months with a wide range (0.3–76.3 months) due to the different survival times of the patients. However, the portion of the patients with a follow-up period more than 2 years was low with 14.4%. The estimated local control rates at 6-, 12-, 24-, 36-, and 72 months were 99.2, 89.0, 67.2, 64.6, and 64.6%, respectively. Nearly half of the lesions achieved partial remission, whereas only 14.8% showed local progression, defined as 20% tumor growth. The most recent report using a LINAC-SRS system analyzed 110 cerebral metastases in 40 patients [18]. In that series, by Gregucci et al., the complete and partial remission rates were comparable to those in our series (complete remission: 10.9% vs. 22.2%; partial remission: 46.4% vs. 46.3%), while the progressive disease rate was lower in their series (1.0% vs. 14.8%) [17]. This discrepancy is most probably due to the different definition of progression; they defined progression as 50% tumor growth [18] compared to our series where it was defined as 20%. Noel et al. reported on approximately 227 metastases in 117 elderly patients also using LINAC-SRS with a local control rate of 91% at 12 months, similar to our series [15]. Minniti et al. also reported the results of a series of 102 elderly patients with 183 metastases treated with LINAC-SRS who attained 84% local control at 12 months (17). Kim et al. analyzed 74 metastases in 44 elderly patients treated with GammaKnife, where local progression was observed in 12% of the patients [16]. Chen et al. compared 37 elderly patients treated with CyberKnife or LINAC-SRS with 82 patients treated by WBRT focusing on toxicity [23]. The authors concluded that WBRT was associated with greater toxicity compared to SRS in elderly and very elderly patients with brain metastases [23]. Gregucci et al. also reported no severe adverse effects worse than grade 2 after SRS [18]. Mininiti et al. reported a neurological complication rate of 13% [17]. In our series, KPS deteriorated by ≥ 10 points only in 1% of the patients, while neurological complications occurred only in 1% of patients. Overall, 14 patients underwent WBRT, which apparently did not cause more significant complications than SRS alone. Evidence regarding the safety of SRS after WBRT was recently reported by Lohkamp et al. [24], and was confirmed by our study findings where even fewer complications were observed.

Although, our study does not provide a head to head comparison of SRS to WBRT, we have to highlight the limitation of SRS in regard to the lack of distant control with our estimated distant progression-free rates. The prospective randomized trial by Brown et al. [5, 25] reported a significantly shorter time to intracranial failure for SRS alone compared with SRS plus WBRT (hazard ratio, 3.6; 95% CI, 2.2–5.9; p < 0.001). Despite a better intracranial tumor control rate associated with WBRT, no improvement in survival time occurred. Importantly, fewer patients underwent salvage therapy after SRS plus WBRT than after SRS alone (7.8% vs 32.4%, respectively; difference, − 24.6%; 95% CI, − 35.7 to − 13.5%; p < 0.001). In our cohort, as a salvage treatment 21.6% (n = 21) of the patients received another CK-SRS, 5.2% (n = 5) of the patients needed a surgery and 3.1% (n = 3) required a WBRT. These were well-tolerated. In regard of the quality of life, physical and social functioning as well as drowsiness and motor dysfunction deteriorated in our small cohort after SRS with complete questionnaires. However, this subjective evaluation was not reflected in the objective complications. Therefore, we assume that older patients who continue to age during multimodal treatment will experience some decrease in quality of life. However, due to the low number of the questionnaires after the SRS our analysis cannot judge this aspect sufficiently. In the series of Brown et al. [5, 25] there was less cognitive deterioration at 3 months after SRS alone (40/63 patients, 63.5%) than when combined with WBRT (44/48 patients, 91.7%; difference, − 28.2%; 90% CI, − 41.9 to − 14.4%; p < 0.001). Quality of life (QoL) was higher at 3 months with SRS alone, including overall quality of life (p = 0.001). A retrospective study by Chen et al. [26] compared SRS and WBRT in 119 geriatric patients with overall 811 lesions (≥ 70 years, ≤ 10 brain metastases). In univariate analysis, fatigue, headache and RTOG CNS toxicity (68.0% vs. 89.0%, p = 0.009) and KPS decline (2.0% vs. 35%. p = 0.0005) was significantly lower in the SRS arm. The multivariate analysis confirmed a higher toxicity after 3 months for the WBRT arm. In contrast to the QUARTZ study, where the combination of WBRT and dexamethasone showed only a small difference in QoL and no difference in OS between the two groups [27], our data with a median survival of 55.6 weeks, preserved KPS and QoL, and no severe acute complications, show a benefit of local therapy with the CK-SRS instead of best supportive care. In light of this accumulated evidence of good local control rates and overall low toxicity rates, SRS should be preferred over WBRT in elderly patients with a certain number of brain metastases despite the limitation for distant tumor control.

As SRS is a purely local therapy, we focused on predictive factors for local control rather than OS and distant progression-free control. The univariate analyses identified older age, female sex, and larger PTV as risk factors for local progression in our cohort, while in the multivariate analysis older age and female sex remained significant, with PTV approximating statistical significance. Importantly, the primary diseases and GPA score did not play a significant role in the regression analysis. We chose the GPA score for our analysis because in two previous studies the score index for radiosurgery (SIR) was superior to the recursive partitioning analysis (RPA) score and in a second analysis the GPA score was the strongest prognostic factor for survival compared to the SIRS [12, 28]. In regard to primary diseases, one possible explanation could be more regular MRI brain scans in some tumor entities such as melanoma in the routine that enables a faster identification of a brain metastases that then response better to a local treatment despite radioresistancy. The prescribed dose did neither play a role in the local control, that is probably due to the overall comparable dose regime in our cohort. Gregucci et al. also identified PTV and BED ≥ 40 Gy as factors associated with local control [18], whereas Minniti et al. could not identify factors predictive of local control [17]. The fact that female sex was identified as a risk factor in our series might be due to the distribution of the primary tumors between the sexes. For instance, breast cancer comprised up to 25% of the primary tumors amongst women and had the worst local control. In this regard, we analyzed the proportion of radioresistant tumors amongst both sexes, however the radiation therapy resistant histologies like malignant melanoma and renal cell carcinoma were balanced between male and female. Triple negative breast cancer was diagnosed in only one woman. One further aspect is the total PTV pro patient that was slightly higher for female patients but did not reach a significant level. Overall, our data did not provide the reason for this observation between female and male patients. We identified older age as a potential risk factor for worse local control; however, Watabene et al. performed a case-matched study comparing SRS applied to very elderly (≥80 years old) patients with that applied to patients between 65 and 79 years old [23] and reported similar local control and complication rates between the groups, signifying that SRS should not be only preserved for patients younger than 80 years [29]. We assume that differences in systemic treatments provided to very elderly patients may explain the effect of age on local control. Although there was no obvious difference in the analysis of the systemic therapy amongst different age groups, we could not supply the exact treatment details. As this was a retrospective study, our data lacked sufficient power to analyze this factor sufficiently.

The major weakness of this study was its retrospective design and the unavoidable preselection of the patients due to the outpatient treatment that requires a reasonable pre-treatment KPS. In any case, our study provides valuable data for this patient group, as it is the largest to date published lesion series concerning elderly patients with brain metastases who were treated using CK.

Radiosurgical treatment of elderly patients with brain metastases could be a safe and efficient option with preserved KPS and QoL. The multivariate analysis shows older age and female sex were predictive factors of local progression, while for the GPA no significance was detected. This analysis provides a basis for further SRS treatment of elderly patients with a limited number of metastases, despite the GPA classification.