Abstract
Background/Aim: Very elderly patients may benefit from individualized treatment. A survival score was created for patients aged 80+ receiving radiosurgery or fractionated stereotactic radiotherapy for 1-2 brain metastases. Patients and Methods: Thirteen patients were retrospectively evaluated. Characteristics showing significant associations with survival or trends were used for analysis. Prognostic groups were calculated from scoring points of these characteristics (0=worse, 1=better survival) added for each patient. Results: Survival was significantly associated with performance score (p=0.010). Trends were found for histology (p=0.066) and diameter of lesions (p=0.071). Three groups were created (0, 1-2, 3 points) with 6-month survival rates of 0%, 56%, and 100% (p=0.025). Positive predictive values (PPVs) to predict death ≤6 months were 100% with the new score vs. not available and 50% with previous scores; PPVs regarding survival ≥6 were 100% vs. 75% and 67%. Conclusion: Given its limitations, the score was more precise than previous tools and can serve for orientation in patients aged 80+.
- Very elderly patients
- brain metastases
- radiosurgery
- fractionated stereotactic radiotherapy
- survival score
Brain metastases occur in 20-40% of adult oncologic patients during the course of their disease (1, 2). A considerable number of these patients receive radiotherapy alone, either as whole-brain irradiation (WBI), as local radiotherapy with single-fraction stereotactic radiosurgery (SRS) or fractionated stereotactic radiation therapy (FSRT), or as a combination of WBI and SRS/FSRT (1-3). Single-fraction SRS is generally limited to lesions with a maximum diameter of 4 cm. Depending on the size of the brain metastases, doses of SRS often range between 15 and 24 Gy (4). Common dose-fractionation regimens of FSRT include 3×9-10 Gy and 5×5-7 Gy. Local radiotherapy is generally used for up to four brain metastases, although it may be applied to selected patients with a greater number of lesions (5). Moreover, since randomized trials have demonstrated that the combination of local radiotherapy and WBI significantly increases the risk of neurocognitive decline, SRS and FSRT are increasingly used without WBI (6, 7).
Since the duration of a treatment session increases with the dose per fraction, sessions of SRS and FSRT take longer than sessions of conventional local radiotherapy with doses per fraction of 2-4 Gy. Moreover, the integrated head mask required for SRS and FSRT is much tighter than the mask used for conventional irradiation. Particularly, very elderly, or frail patients may not be able to tolerate the longer lasting treatment sessions of SRS/FSRT and the tight mask and receive short-course conventional radiotherapy instead. However, due to technical improvements, SRS and FSRT will likely be used more frequently also for these two groups. Very elderly patients, often defined as 80+, are considered a separate group of patients due to the higher comorbidity index and reduced organ function when compared to other age groups (8-11). These patients would benefit particularly from individualized treatments to avoid over- or undertreatment. When designing an individualized treatment program, several aspects need to be considered including the patient’s survival prognosis, which can be estimated with the help of survival scores. Several scores are already available for patients receiving SRS or FSRT for brain metastases but were not designed specifically for very elderly patients (12-17). To provide the optimal individualized treatment, a separate score should be available for this age group. In the present study, we have developed the first specific survival score for very elderly patients (80+) with very few brain metastases assigned to SRS or FSRT alone. In addition, the new score was compared to two previous tools developed for elderly patients (65+) (16, 17).
Patients and Methods
Thirteen very elderly patients (80+) who received single-fraction SRS or FSRT for one or two brain metastases between 1999 and 2020 were included in this retrospective study. The study was approved by the Ethics Committee of the University of Lübeck, Germany (file number 22-062). SRS (n=5) was performed with 1×20 Gy, and FSRT with 3×10-11 Gy (n=5), 5×5-6 Gy (n=2), or 7×5 Gy (n=1). Treatment was performed with a conventional linear accelerator (Varian Medical Systems, Palo Alto, CA, USA) (n=10) or a CyberKnife® system (Accuray Inc., Sunnyvale, CA, USA) (n=3). The type of radiotherapy (SRS vs. FSRT), the period of treatment (until 2014 vs. 2015-2020; median number of patients treated until 2014) and eight additional characteristics were evaluated for associations with survival. These characteristics included age (≤83 vs. ≥84 years, median age=83 years), sex (female vs. male), Karnofsky performance score (70-80 vs. ≥90), primary tumor type (lung cancer vs. melanoma vs. other types), number of cerebral lesions (1 vs. 2), maximum cumulative diameter of lesions (≤16 vs. >16 mm, median diameter=16 mm), infratentorial involvement (no vs. yes), and extracranial metastases (no vs. yes).
Survival was calculated from the day of SRS or the last day of FSRT. The survival analyses were performed using the Kaplan–Meier method and the log-rank test. Characteristics that showed a significant association with survival (p<0.05) or a trend (p<0.08) were used for creating the survival score. Zero points were assigned to the subgroup with worse and one point to the subgroup with more favorable survival outcomes, respectively. For each patient, the scoring points of the corresponding characteristics were added to obtain individual patient scores. Considering the 6-month survival rates of the resulting patient scores, three groups were designed, representing patients with poor, intermediate, and favorable survival prognoses, respectively.
In addition, the new survival score was compared to two previous tools that were developed in a cohort of elderly patients (65+) receiving SRS or FSRT for very few brain metastases (16, 17). Diagnostic accuracy of both scores was assessed by calculating the positive predictive values (PPVs) to correctly predict death within 6 months (worst prognostic groups) and survival for at least 6 months (best prognostic groups). The PPVs for prediction of death were calculated as follows:
PPV=[patients dying/(patients dying + patients not dying)] ×100. The PPVs for prediction of survival were calculated as follows: PPV=[patients surviving/(patients surviving + patients not surviving)] ×100.
Results
According to the survival analyses, a KPS of ≥90 was significantly associated with better outcomes (p=0.010). In addition, trends for improved survival were found for favorable primary tumor types (melanoma and lung cancer; p=0.066) and maximum cumulative diameter of brain metastases ≤16 mm (p=0.071). The results of the complete survival analyses are summarized in Table I.
Survival rates at 3, 6, and 12 months following radiosurgery or fractionated stereotactic radiotherapy.
When adding the scoring points of KPS, primary tumor type, and maximum cumulative diameter for each patient, resulting patient scores ranged between 0 and 3 points (Table II). Corresponding 6-month survival rates were 0% (0 points), 40% (1 point), 75% (2 points), and 100% (3 points), respectively (p=0.019). When patients who achieved 1 or 2 points were combined to one group, three prognostic groups were obtained, i.e., 0 points (n=1), 1-2 points (n=9) and 3 points (n=3). Median survival times of these groups were not available, 8 months, and >20 months (not reached until 20 months), respectively, and 6-month survival rates were 0%, 56%, and 100%, respectively (p=0.025, Figure 1).
Characteristics used for the survival score and corresponding scoring points.
Using the new score: Kaplan–Meier curves of the three prognostic groups 0 points, 1-2 points, and 3 points. The p-value was calculated with the log-rank test.
When applying the first previous survival score developed for elderly patients, the Yamamoto-Score, no patient of the present cohort had 0 or 1 point. Median survival times for the other two groups were 4 months (2-3 points) and 12 months (4-5 points), respectively, and 6-month survival rates were 40% and 75%, respectively (p=0.066, Figure 2) (16). When using the second previous tool, i.e., our previous score, median survival times were 6 months (12-14 points), 8 months (15-16 points) and 11 months (17-19 points), respectively, and 6-month survival rates were 50%, 60%, and 67%, respectively (p=0.60, Figure 3) (17). PPVs for correct prediction of death within 6 months were 100% with the new score vs. 50% with the second of the previous scores, respectively (17). PPVs for correct prediction of survival for at least 6 months were 100% vs. 75% (Yamamoto-Score) and 67% (our previous score), respectively (16, 17).
Using the Yamamoto-Score (16): Kaplan–Meier curves of the three prognostic groups 0-1 points, 2-3 points, and 4-5 points. The p-value was calculated with the log-rank test.
Using our previous score (17): Kaplan–Meier curves of the three prognostic groups 12-14 points, 15-16 points, and 17-19 points. The p-value was calculated with the log-rank test.
Discussion
In the field of oncology, it is widely accepted to consider elderly patients (aged 65+) a separate group. Due to demographic changes, the number of these patients is constantly increasing (18, 19). This applies also to the group of very elderly patients, generally defined as aged 80+ (8-11). For both groups, personalization of their treatment is important to achieve optimal outcomes. This holds true for all oncologic patients but particularly for elderly and very elderly patients, who are generally frailer, compromised by other diseases, and have decreased function of organs and bone marrow, when compared to patients at younger age (20). The process of treatment personalization should consider several aspects including the patient’s remaining lifespan. Therefore, it is important to be able to estimate an individual patient’s prognosis prior to the start of treatment. For this purpose, physicians can be supported by survival scores. Since metastatic disease can differ considerably regarding sites, numbers, and biology of the lesions, these patients may benefit particularly from personalized treatment. These considerations include elderly and very elderly patients with brain metastases. Several survival scores have been developed for elderly patients with brain metastases aged 65+ (16, 17, 21-29). However, no score has been available so far for very elderly patients. In the present study, a survival score has been created specifically for patients aged 80+, who received SRS or FSRT for very few brain metastases. Based on the prognostic factors, Karnofsky performance score, primary tumor type, and cumulative diameter of brain metastases, three prognostic groups have been designed.
In the 0 points group (poor prognosis), death occurred within 3 months of radiotherapy. Therefore, these patients may be considered for best supportive care or a short course (e.g., 5×4 Gy over 1 week) of conventional local radiotherapy, which would take less time per treatment session and not require a very tight head mask used for SRS or FSRT. Patients of the 1-2 points group (intermediate prognosis) had a median survival time of 8 months and a 6-month survival rate of 56%. These patients may live long enough to experience late side effects of the radiation treatment. Therefore, brain metastases >2 cm should be treated with FSRT instead of single-fraction SRS. In the retrospective study of Minniti et al. that compared SRS and FSRT in 289 patients with 343 brain metastases, FSRT (3×9 Gy) resulted in better 1-year local control (91% vs. 77%, p=0.01) and a reduced risk of 1-year radio-necrosis (9% vs. 18%, p=0.01) than SRS with 1×15-18 Gy (30). Moreover, concurrent or sequential administration of systemic treatment in addition to SRS or FSRT increases the risk of radio-necrosis (31, 32). The risk will likely be lower after FSRT, since the biological effective dose with respect to late radiation effects is lower with FSRT compared to SRS with a single very high dose (33, 34). These considerations are even more important for the 3 points group of the new score, since these patients lived longer than 20 months following radiotherapy.
When considering these recommendations, the limitations of the present study, mainly the retrospective design and the very small sample size, should be kept in mind. Given these limitations, the present new score merely serves for coarse orientation. However, when comparing the new score to two previous tools developed for elderly patients (65+) receiving SRS or FSRT for a limited number of brain metastases, the new score was more accurate than both previous scores in predicting survival for at least 6 months and more accurate than our previous score in predicting death within 6 months (16, 17).
In summary, given its limitations, the new score was more precise than previous tools and, therefore, can serve for orientation when aiming to personalize treatment for a limited number of brain metastases in patients aged 80+. Validation of the score in a larger cohort of patients is warranted. Since the number of very elderly patients and the use of SRS or FSRT in this age group will increase, such a validation can be expected over the coming years.
Acknowledgements
The study was funded by the European Regional Development Fund through the Interreg Deutschland-Danmark program as part of the project TreaT (148-1.1-21).
Footnotes
Authors’ Contributions
D.R., C.D., S.E.S. and S.J. participated in the design of the study. D.R., C.D. and O.B. provided the data, which were analyzed by S.E.S. The article being drafted by D.R. and S.E.S. was reviewed and approved by all Authors.
Conflicts of Interest
On behalf of all Authors, the corresponding Author states that there are no conflicts of interest related to this study.
- Received September 15, 2022.
- Revision received September 22, 2022.
- Accepted September 23, 2022.
- Copyright © 2022 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).









