Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Support Care Cancer (2013) 21:2671–2678 DOI 10.1007/s00520-013-1840-5 ORIGINAL ARTICLE Best supportive care in patients with brain metastases and adverse prognostic factors: development of improved decision aids Carsten Nieder & Jan Norum & Astrid Dalhaug & Gro Aandahl & Kirsten Engljähringer Received: 31 January 2013 / Accepted: 25 April 2013 / Published online: 18 May 2013 # Springer-Verlag Berlin Heidelberg 2013 Abstract Purpose This study aimed to develop a survival prediction model that might aid decision making when choosing between best supportive care (BSC) and brain radiotherapy (RT) for patients with brain metastases and limited survival expectation. Methods A retrospective analysis of 124 patients treated with BSC, whole brain radiotherapy (WBRT), or radiosurgery was conducted. All patients had adverse prognostic features defined as 0–1.5 points according to the diagnosis-specific graded prognostic assessment score (DS-GPA) or GPA if primary tumor type was not among those represented in DS-GPA. Kaplan–Meier survival curves were compared between patients treated with BSC or RT in different scenarios, reflecting more or less rigorous definitions of poor prognosis. If survival was indistinguishable and this result could be confirmed in multivariate analysis, BSC was considered appropriate. Results Irrespective of point sum examined, DS-GPA by itself was not a satisfactory selection parameter. However, we defined a subgroup of 63 patients (51 %) with short survival irrespective of management approach (only 5 % of irradiated patients survived beyond 6 months; they had newly C. Nieder (*) : A. Dalhaug : G. Aandahl : K. Engljähringer Department of Oncology and Palliative Medicine, Nordland Hospital, 8092 Bodø, Norway e-mail: [email protected] C. Nieder : J. Norum : A. Dalhaug Institute of Clinical Medicine, Faculty of Health Sciences, University of Tromsø, 9037 Tromsø, Norway J. Norum Department of Oncology, University Hospital of North Norway, 9038 Tromsø, Norway J. Norum Northern Norway Regional Health Authority Trust, Bodø, Norway diagnosed, treatment-naïve lung cancer), i.e., patients in whom foregoing RT was unlikely to compromise survival. These were patients with 0–1.5 points and aged ≥75 years, had Karnofsky performance status ≤50, or had uncontrolled primary tumor with extracranial metastases to at least two organs. Conclusions BSC is a reasonable choice in patients with limited life expectancy. After successful external validation of the selection criteria developed in this analysis, identification of patients who are unlikely to benefit from WBRT might be improved. Keywords Brain metastases . Prognostic factors . Radiotherapy . Best supportive care Background Recent attempts towards personalized cancer treatment are expected to improve individual patient’s outcome and limit the rapid growth of health care budgets. Development of prediction tools might facilitate informed decisions about which treatment is likely to work or to be futile. Avoiding futile treatment efforts saves patients from unnecessary toxicity, travelling, and time spent in hospitals or oncology practices [1]. Several studies have shown that patients with noncurable metastatic cancer might receive too aggressive and long-standing treatment during the last months of life. The focus on optimal supportive care may be lost or prioritized too late. Many patients with brain metastases belong to this group [2, 3]. These considerations led to the activation of a prospective randomized study of best supportive care (BSC) vs. whole brain radiotherapy (WBRT) known as the QUARTZ trial [4]. Inclusion was limited to patients with inoperable brain metastases from non-small cell lung cancer 2672 (NSCLC) for whom the clinician and patient is uncertain of the benefit of WBRT. Interim analysis of the results suggested comparable quality-adjusted life years (QALY) and survival gain in the two arms. Since many patients with primaries other than NSCLC face the same difficult dilemma, we recently compared BSC to WBRT in patients with different primary tumors and poor prognosis where uncertainty about the benefit of WBRT existed [5]. Our findings confirmed that BSC did not shorten survival and was an appropriate choice for the majority of these patients. We concluded that it would be desirable to develop objective selection criteria which allow for prediction of whether or not a given patient can safely forego WBRT. Not all clinicians are comfortable basing important treatment decisions upon subjective statements such as “when the clinician and patient is uncertain of the benefit of WBRT.” For this purpose, many clinicians prefer to rely on well-defined prognostic factors. Attempts have been made to predict survival by the use of prognostic scores such as the Radiation Therapy Oncology Group (RTOG)’s recursive partitioning analysis (RPA) classes [6]. More recently, the RTOG proposed a new index, the graded prognostic assessment (GPA) [7], which was followed by the diagnosis-specific GPA (DS-GPA) scores for patients with renal cell carcinoma, malignant melanoma, lung, breast, and gastrointestinal cancers [8]. We were able to confirm that each of these scores is valid in real-world patients treated outside clinical trials [9–11]. Nevertheless, substantial variations of actual survival make it impossible to rely solely on RPA, GPA, or even a combination of GPA and two other scores when deciding about management of individual patients with adverse prognostic features [12]. In extension of our previous work that was performed before the DS-GPA became available, we aimed at the development of less error-prone survival prediction models for patients who might be candidates for BSC. We hypothesized that the DS-GPA might be a better fundament than the previously tested scores. Methods For this retrospective single-institution study, consecutive patients treated during the 8-year time period between 2005 and 2012 were chosen. All patients had adverse prognostic features, arbitrarily defined as 0–1.5 points according to the definitions of the DS-GPA score. The GPA score was used in patients with primary tumors not covered by the DS-GPA because it was previously shown that GPA is a pragmatic and valid choice [13]. All 124 patients were derived from a previously described brain metastases database, which is maintained and updated at the first author’s institution in Bodø [14, 15]. All patients with brain metastases presented Support Care Cancer (2013) 21:2671–2678 to this institution were registered in the database, regardless of prognosis or management approach (BSC or radiotherapy [RT]). Irradiated patients had immediate WBRT or stereotactic radiosurgery (SRS) without preceding systemic therapy after the diagnosis of brain metastases. None of the patients had previous brain RT (prophylactic or other). A shared decision towards treatment approach was made by the patient, caregiver(s), and treating physician(s), taking into account the life expectancy, local and systemic disease extent, and patient preferences (no standardized use of any particular prognostic model). Six patients (5 %) were alive at date of last follow-up (November 1, 2012) and thus censored. Their median follow-up was 4.9 months (range, 2.6–12.4 months). Date of death was known in all other patients. We analyzed survival outcomes and prognostic factors (baseline and treatmentrelated). For the comparison of dichotomous variables, the chi-square test and Fisher’s exact test, where applicable, were employed, and for continuous variables, the Mann–Whitney U test was used. Survival from imaging diagnosis of brain metastases was calculated with the Kaplan–Meier method and compared between different groups (BSC vs. active treatment) with the log-rank test. For multivariate prediction of survival, Cox regression analysis was used (forward stepwise data selection method). A p value ≤0.05 was considered statistically significant. Results Table 1 shows detailed information on all 124 patients with adverse prognostic features, i.e., 0–1.5 points according to the DS-GPA (GPA if DS-GPA was not applicable, n=7) in this study. As this was an intention-to-treat analysis, we included patients who failed to complete their course of WBRT. Initially, we tested in all 124 patients whether or not active treatment improved survival compared to BSC. We found that actuarial median survival from imaging diagnosis was longer in each of the RT arms compared to BSC (BSC, 1.5 months; WBRT 20 Gy, 2.2 months (p=0.07); WBRT 30 Gy, 3.0 months (p=0.0001); WBRT with or without boost to doses higher than 30 Gy, 5.7 months (p=0.01); SRS, 5.3 months (p=0.009); note that all patients with incomplete RT always were included in the appropriate group). Besides median survival, survival beyond 6 months was also improved (Fig. 1). Multivariate analysis with treatment arms (p=0.001), RPA classes (p=0.08), and DS-GPA (or GPA in seven patients; p=0.04)) confirmed that survival after BSC was not equivalent to that after RT. A second multivariate model included individual prognostic factors (age, number of brain metastases, Karnofsky performance status (KPS), presence of extracranial metastases, primary tumor control, and gender) in addition to those included in Support Care Cancer (2013) 21:2671–2678 2673 Table 1 Patient characteristics (n=124) Parameter Number Percent Treatment BSC WBRT 20 Gy intended dose WBRT 30 Gy intended dose WBRT±boost>30 Gy SRS Incomplete RT course Female patients Male patients Small cell lung cancer NSCLC Breast cancer Malignant melanoma Renal cell carcinoma Gastrointestinal cancer Unknown primary Other 33 28 54 5 4 9 64 60 16 54 10 13 7 17 3 4 27 23 44 4 3 7 52 48 13 44 8 10 6 14 2 3 67 57 11 113 62 14 40 70 106 18 1 40 83 66 (40–88) 60 (30–90) 54 46 9 91 50 11 32 57 85 15 1 32 67 Controlled primary tumor Uncontrolled primary tumor Extracranial metastases absent Extracranial metastases present At least two organs involved Single brain metastasis Two or three brain metastases More than three brain metastases DS-GPA poor DS-GPA intermediate–poor RPA class 1 RPA class 2 RPA class 3 Median age, years (range) Median KPS, range Median intervala, months (range) Median number of brain metastases Median DS-GPA (or GPA) 6 (0–120) 4 (1–51) 0.5 (0–1.5) BSC best supportive care, WBRT whole brain radiotherapy, SRS stereotactic radiosurgery, DS-GPA diagnosis-specific graded prognostic assessment (0–1 point, poor; 1.5–2 points, intermediate–poor; 2.5–3 points, intermediate–good), RPA recursive partitioning analysis, KPS Karnofsky performance status a Interval from initial cancer diagnosis to brain metastases the first model. Again, survival was not equivalent (p=0.01 for treatment arms, p=0.03 for DS-GPA/GPA, and p=0.05 for KPS). In addition to being statistically significant (with one exception of p=0.07), the survival improvement in the RT arms appeared clinically meaningful because it was longer than the time spent on RT, even for the 20-Gy WBRT group where the survival difference was minor (0.7 months). Based on these findings, we cannot recommend the use of DS-GPA/GPA 0–1.5 points as selection criterion for BSC. Next we tested whether or not restricting DS-GPA/GPA to 0–1 points would be more appropriate. This analysis followed the same path and included 106 patients. Based on the almost identical survival differences observed (BSC, 1.5 months; WBRT 20 Gy, 1.9 months; WBRT 30 Gy, 2.7 months; WBRT with or without boost to doses higher than 30 Gy, 5.7 months; SRS, 5.3 months) and confirmatory multivariate analyses identical to those described previously, we rejected the use of this criterion too. Then, we chose to restrict DS-GPA/GPA to 0–0.5 points (n =68). Again, significant differences were seen (BSC, 0.9 months; WBRT 20 Gy, 2.2 months; WBRT 30 Gy, 2.1 months; no patients with other RT approaches), p=0.049 and p=0.005, respectively, which were confirmed in multivariate analyses as described previously. Even limiting the analysis to patients with 0 points only (n=40) did not result in satisfactory survival differences, i.e., less than the 2 weeks spent on a 30-Gy WBRT regimen. Median actuarial survival was 0.8 months (BSC), 1.7 months (20 Gy WBRT), and 2.1 months (30 Gy WBRT), p=0.03 and p=0.001, respectively. Given that DS-GPA/GPA alone was not sufficient to decide that withholding WBRT would be appropriate because survival would not be compromised, we hypothesized that adding other factors to this validated prognostic score could improve prediction of short survival. We looked at age, KPS, primary tumor control, and extracranial metastases, i.e., established factors forming the basis of the RPA classes, and combinations of these factors such as uncontrolled primary tumor plus extracranial metastases or metastases to at least two organs (liver plus lung, bone plus adrenal gland, etc.). For age and KPS, different cutoffs were analyzed in order to optimize the performance of each parameter. In each case, we compared Kaplan–Meier curves for three groups of patients with DS-GPA/GPA 0–1.5 points plus one additional adverse factor, e.g., KPS <50, treated either with BSC or RT (WBRT 20 Gy vs. all other types and doses combined, termed more aggressive RT). After optimization of cutoff and parameter combinations, we found identical survival curves for patients managed with BSC or RT who fulfilled the following criteria: DS-GPA/GPA 0–1.5 points and age at least 75 years, DS-GPA/GPA 0–1.5 points and KPS ≤50, and DS-GPA/GPA 0–1.5 points and uncontrolled primary tumor with extracranial metastases to at least two organs. The performance of our survival prediction model did not improve by altering the DS-GPA/GPA requirement, e.g., limiting inclusion to 0–1 or 0–0.5 points. Finally, we grouped together all patients from the three subgroups with DS-GPA/GPA 0–1.5 points and age at least 75 years or KPS ≤50 or uncontrolled primary tumor with extracranial metastases to at least two organs, i.e., those who are unlikely to experience improved survival from RT. 2674 Support Care Cancer (2013) 21:2671–2678 Fig. 1 Actuarial Kaplan–Meier survival curves for patients treated with short-course WBRT (20 Gy intended dose) or more aggressive RT (intention to treat) or BSC; differences were statistically significant, except for BSC vs. 20 Gy WBRT (p=0.07), logrank test Diagnosis-specific graded prognostic assessment (DS-GPA) 0-1.5 points in all patients Table 2 shows the characteristics of these 63 patients (51 %). Median survival was 1.5 months (BSC), 1.7 months (WBRT 20 Gy), and 2.1 months (more aggressive RT), p=0.87, p=0.24, and p=0.21, respectively (note that significantly more patients who had aggressive RT belonged to RPA class 2 (Table 2), which explains most of the survival difference seen here). Multivariate analysis confirmed that treatment arm was not significantly associated with survival (p=0.53, further variables included were DS-GPA/GPA with p=0.05, gender with p=0.2, age with p=0.36, RPA with p=0.9, and KPS with p=0.98). The survival curves are shown in Fig. 2. In contrast to our first, purely DS-GPA/GPA-based analyses, which gave unsatisfactory results with both significantly better median survival and considerable proportions of irradiated patients who survived for more than 6 months, very few irradiated patients survived for more than 6 months (0 out of 11 with age 75 years or older, 1 out of 41 with KPS ≤50, and 1 out of 25 with uncontrolled primary tumor and extracranial metastases to at least two organs; some patients fulfilled two or three criteria at the same time). Thus, only 2 out of 41 irradiated patients (5 %, both had newly diagnosed untreated NSCLC) survived for more than 6 months. In other words, applying these criteria to recommend BSC has a very low risk of compromising patients’ survival. The last step consisted of comparing these selection criteria to those identified in our previous, smaller analysis from the year 2010 [12]. At that time, we had relied on older scores and recommended that patients with KPS <80, uncontrolled primary tumor, presence of extracranial metastases, and no intent to treat systemically might best be managed with BSC. We selected all patients fulfilling these four criteria from our database. Among 37 patients, 36 also fulfilled the criterion of DS-GPA/GPA 0–1.5 points, while 1 patient scored 2 points. Eighteen patients were managed with BSC, 8 with WBRT 20 Gy intended dose, and 11 with WBRT 30 Gy intended dose. Median survival was 1.8 months after BSC, 1.9 months after short-course WBRT (p = 0.89), and 3.0 months after longer-course WBRT (p=0.28). Figure 3 shows the survival curves. It should be noted that RT did not result in long-term survival. Imbalances in baseline prognostic factors favored the longer-course WBRT group. Multivariate analysis, including baseline prognostic factors according to the methods described previously, did not confirm that treatment arm independently influenced survival. A disadvantage of these 4 selection criteria is that only 37 patients would be considered appropriate for BSC, while the new model would recommend BSC for 63 patients. Discussion There is increasing awareness and acceptance of terminal care strategies for patients with metastatic cancer, which avoid overtreatment and rather focus on optimal palliation. Evidence from several sources, including interim data from a randomized trial, suggest that avoiding RT and providing BSC for patients with brain metastases and adverse prognostic features does not shorten survival [4, 5]. The already Support Care Cancer (2013) 21:2671–2678 Table 2 Characteristics of patients (n=63) considered appropriate candidates for BSC (n=22 managed with BSC, 18 who received short-course WBRT with 20 Gy intended dose, and 23 who received more aggressive RT) Note that patients who received more aggressive RT had more favorable baseline prognostic factors, including younger median age, better median KPS, and higher proportion of RPA class 2 WBRT whole brain radiotherapy, DS-GPA diagnosis-specific graded prognostic assessment (0–1 point, poor; 1.5–2 points, intermediate–poor), RPA recursive partitioning analysis, KPS Karnofsky performance status a Interval from initial cancer diagnosis to brain metastases Parameter 2675 BSC 20 Gy Number Percent Female patients Male patients Small cell lung cancer NSCLC Breast cancer Malignant melanoma Renal cell carcinoma Gastrointestinal cancer Other Controlled primary tumor Uncontrolled primary tumor Extracranial metastases absent Extracranial metastases present Single brain metastasis Two or three brain metastases More than three brain metastases DS-GPA poor DS-GPA intermediate–poor RPA class 3 12 10 3 10 1 5 1 0 2 9 13 3 19 1 7 14 22 0 22 55 45 14 45 5 23 5 0 9 41 59 14 86 5 32 64 100 0 100 RPA class 2 Median age, years Median KPS Median intervala, months 0 75 5 0 Incomplete RT course >20 Gy Number Percent Number Percent 1 6 6 26 9 9 2 10 0 0 0 3 3 9 9 1 17 3 4 11 17 1 17 50 50 11 56 0 0 0 17 17 50 50 6 94 17 22 61 94 6 94 16 7 1 11 3 2 3 1 2 5 18 2 21 1 8 14 21 2 13 70 30 4 48 13 9 13 4 9 22 78 9 91 4 35 61 91 9 57 1 68 5 6 10 67 6 43 Median number of brain metastases 6 4 5 4 2 5 Median DS-GPA (or GPA) 0 0.5 0.5 mentioned QUARTZ trial also suggested that comparable QALY results can be obtained [4]. Ideally, unequivocal objective selection criteria such as RPA class 3 or low DSGPA score would allow for straightforward decision making when clinicians consider referral for palliative RT. However, as shown in our previous studies, prediction of short survival is more complex [9–12]. In the absence of generally accepted tools or decision aids, the QUARTZ trial included patients for whom the clinician and patient was uncertain of the benefit of WBRT. With the present study, we attempted to develop a more objective model for patient selection. For at least two to three decades, researchers have attempted to predict survival of patients with brain metastases [6–8, 16–18]. Several score instruments have been developed and compared, resulting in the identification of subgroups with limited median survival, yet difficult to predict individual outcome. In other words, even the most unfavorable group contained patients with very short, intermediate, and longer survival. We assumed that the recently published and validated DS-GPA score [8, 11, 19] might be a better tool than previous scores. We decided not to limit our study to patients with 0–1 points, which comprise the most unfavorable prognostic group as originally defined by Sperduto et al. [8], because, in our experience, those with 1.5 points had only marginally better survival [11]. Our retrospective intention-to-treat analysis was designed to prevent bias to the largest possible degree and included all patients managed with BSC and different types of RT at our institution. However, the methodological limitations of retrospective studies must be acknowledged. We cannot exclude that our strategy of optimizing prognostic models as far as possible and looking at a large number of potential variables and cutoffs might have resulted in overfitting of data. Our database was not large enough to confirm our results in a validation group. It is, therefore, necessary to validate the results externally in other databases before recommending widespread clinical use. We also acknowledge that survival is only one aspect when patients with advanced cancer are diagnosed with brain metastases and that symptom burden and quality of life should be considered when recommending treatment strategies [20–23]. Individual patients might have variable conceptions of what 2676 Support Care Cancer (2013) 21:2671–2678 Fig. 2 Actuarial Kaplan–Meier survival curves for patients treated with short-course WBRT (20 Gy intended dose) or more aggressive RT (intention to treat) or BSC; differences were not significant, log-rank test Diagnosis-specific graded prognostic assessment (DS-GPA) 0-1.5 points, age 75 years or older or KPS <=50 or uncontrolled primary tumor with extracranial metastases to at least 2 organs can be considered a meaningful prolongation of survival. From our point of view, minimal prolongation of survival by 2–3 weeks does not justify additional burden and toxicity during the terminal phase of disease (notably if the gain equals the time spent on active RT). We cannot recommend the use of DS-GPA alone for the purpose of our study because such a strategy could result in undertreatment and compromised survival. However, patients with DS-GPA score 0–1.5 points and one or more of three additional features might be appropriate candidates for BSC (age at least 75 years, KPS ≤50, or uncontrolled primary tumor with extracranial metastases to at least two organs). We were able to confirm that a previous selection model is valid too (KPS <80, uncontrolled primary tumor, Fig. 3 Actuarial Kaplan–Meier survival curves for patients treated with short-course WBRT (20 Gy intended dose) or longer-course WBRT (30 Gy intended dose) or BSC; differences were not significant, log-rank test KPS <80, uncontrolled primary tumor with extracranial metastases, no intent to treat systemically Support Care Cancer (2013) 21:2671–2678 presence of extracranial metastases, and no intent to treat systemically), but fewer patients were assigned to BSC when using this model. Lagerwaard et al. reported on 118 patients (9 % of their total patient population with different primary tumors) who were treated with BSC after the diagnosis of brain metastases [17]. Median survival was 1.3 months, and the 6-month survival rate was 4 %. They defined a poor prognosis group with reduced performance status (ECOG 2 or 3), little or no response to steroids, and limited or extensive systemic tumor activity. These patients, about 10 % of patients in their series, had a median survival of only 1.3 months after palliative RT, leading the authors to conclude that a conservative approach in the treatment of this group of patients seems justified. We were not able to examine these criteria in our database due to a lack of recorded steroid response. A large Australian series included 327 melanoma patients managed with BSC between 1952 and 1984 and another 210 patients from the era 1985–2000 [24]. Median survival was 1.7 months (range, 0.8–3.4) in patients treated in the earlier part of the study and 2.1 months (range, 0.9–5.0) in those treated after 1984. This group reported a hazard ratio of 0.85 (95 % confidence interval, 0.7–1.04; p=0.11) for the survival comparison between brain RT and BSC (1985–2000; 210 and 236 patients, respectively). During the time period 1996–2000, 26 % of patients were managed with BSC. Predictive factors were not analyzed in this study. Conclusions Increasing evidence suggests that BSC is appropriate in patients with brain metastases and limited life expectancy. The predictive model presented in our study should be externally validated. It might facilitate the identification of patients who are unlikely to benefit from WBRT. Conflict of interest The authors declare that they have no conflict of interest. References 1. Craighead PS, Chan A (2012) Defining treatment for brain metastases patients: nihilism versus optimism. Support Care Cancer 20:279–285 2. Danielson B, Fairchild A (2012) Beyond palliative radiotherapy: a pilot multidisciplinary brain metastases clinic. Support Care Cancer 20:773–781 3. Yamanaka R, Koga H, Yamamoto Y, Yamada S, Sano T, Fukushige T (2011) Characteristics of patients with brain metastases from lung cancer in a palliative care center. Support Care Cancer 19:467–473 2677 4. Langley RE, Stephens RJ, Nankivell M, Pugh C, Moore B, Navani N, Wilson P, Faivre-Finn C, Barton R, Parmar M, Mulvenna PM, Investigators QUARTZ (2013) Interim data from the Medical Research Council QUARTZ trial: does whole brain radiotherapy affect the survival and quality of life of patients with brain metastases from non-small cell lung cancer? Clin Oncol (R Coll Radiol) 25:e23–e30 5. Nieder C, Norum J, Dalhaug A, Aandahl G, Pawinski A (2013) Radiotherapy versus best supportive care in patients with brain metastases and adverse prognostic factors. Clin Exp Metastasis doi:10.1007/s10585-013-9573-x 6. Gaspar L, Scott C, Rotman M, Asbell S, Phillips T, Wasserman T, McKenna WG, Byhart R (1997) Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys 37:745–751 7. Sperduto PW, Berkey B, Gaspar LE, Mehta M, Curran W (2008) A new prognostic index and comparison to three other indices for patients with brain metastases: an analysis of 1,960 patients in the RTOG database. Int J Radiat Oncol Biol Phys 70:510–514 8. Sperduto PW, Chao ST, Sneed PK, Luo X, Suh J, Roberge D, Bhatt A, Jensen AW, Brown PD, Shih H, Kirkpatrick J, Schwer A, Gaspar LE, Fiveash JB, Chiang V, Knisely J, Sperduto CM, Mehta M (2010) Diagnosis-specific prognostic factors, indexes, and treatment outcomes for patients with newly diagnosed brain metastases: a multi-institutional analysis of 4,259 patients. Int J Radiat Oncol Biol Phys 77:655–661 9. Nieder C, Nestle U, Motaref B, Walter K, Niewald M, Schnabel K (2000) Prognostic factors in brain metastases: should patients be selected for aggressive treatment according to recursive partitioning analysis (RPA) classes? Int J Radiat Oncol Biol Phys 46:297–302 10. Nieder C, Marienhagen K, Geinitz H, Molls M (2009) Validation of the graded prognostic assessment index for patients with brain metastases. Acta Oncol 48:457–459 11. Nieder C, Andratschke NH, Geinitz H, Grosu AL (2012) Diagnosis-specific graded prognostic assessment score is valid in patients with brain metastases treated in routine clinical practice in two European countries. Med Sci Monit 18:CR450–CR455 12. Nieder C, Pawinski A, Molls M (2010) Prediction of short survival in patients with brain metastases based on three different scores: a role for ‘triple-negative’ status? Clin Oncol (R Coll Radiol) 22:65–69 13. Nieder C, Andratschke NH, Geinitz H, Grosu AL (2012) Use of the Graded Prognostic Assessment (GPA) score in patients with brain metastases from primary tumours not represented in the diagnosis-specific GPA studies. Strahlenther Onkol 188:692–695 14. Nieder C, Spanne O, Mehta MP, Grosu AL, Geinitz H (2010) Presentation, patterns of care, and survival in patients with brain metastases: what has changed in the last 20 years? Cancer 117:2505–2512 15. Nieder C, Norum J, Stemland JG, Dalhaug A (2010) Resource utilization in patients with brain metastases managed with best supportive care, radiotherapy and/or surgical resection: a Markov analysis. Oncology 78:348–355 16. Rades D, Evers JN, Veninga T, Stalpers LJ, Lohynska R, Schild SE (2011) Shorter-course whole-brain radiotherapy for brain metastases in elderly patients. Int J Radiat Oncol Biol Phys 81:e467–e473 17. Lagerwaard FJ, Levendag PC, Nowak PJ, Eijkenboom WM, Hanssens PE, Schmitz PI (1999) Identification of prognostic factors in patients with brain metastases: a review of 1292 patients. Int J Radiat Oncol Biol Phys 43:795–803 18. Rades D, Dziggel L, Haatanen T, Veninga T, Lohynska R, Dunst J, Schild SE (2011) Scoring systems to estimate intracerebral control and survival rates of patients irradiated for brain metastases. Int J Radiat Oncol Biol Phys 80:1122–1127 19. Sperduto PW, Kased N, Roberge D, Xu Z, Shanley R, Luo X, Sneed PK, Chao ST, Weil RJ, Suh J, Bhatt A, Jensen AW, Brown 2678 PD, Shih HA, Kirkpatrick J, Gaspar LE, Fiveash JB, Chiang V, Knisely JP, Sperduto CM, Lin N, Mehta M (2012) Summary report on the graded prognostic assessment: an accurate and facile diagnosis-specific tool to estimate survival for patients with brain metastases. J Clin Oncol 30:419–425 20. Khan L, Cramarossa G, Lemke M, Nguyen J, Zhang L, Chen E, Chow E (2013) Symptom clusters using the Spitzer quality of life index in patients with brain metastases—a reanalysis comparing different statistical methods. Support Care Cancer 21:467–473 21. Soffietti R, Kocher M, Abacioglu UM, Villa S, Fauchon F, Baumert BG, Fariselli L, Tzuk-Shina T, Kortmann RD, Carrie C, Ben Hassel M, Kouri M, Valeinis E, van den Berge D, Mueller RP, Tridello G, Collette L, Bottomley A (2013) A European Organisation for Research and Treatment of Cancer phase III trial of adjuvant whole-brain radiotherapy versus observation in patients with one to three brain metastases from solid tumors after Support Care Cancer (2013) 21:2671–2678 surgical resection or radiosurgery: quality-of-life results. J Clin Oncol 31:65–72 22. Steinmann D, Vordermark D, Geinitz H, Aschoff R, Bayerl A, Gerstein J, Hipp M, van Oorschot B, Wypior HJ, Schäfer C (2013) Proxy assessment of patients before and after radiotherapy for brain metastases: results of a prospective study using the DEGRO brain module. Strahlenther Onkol 189:47–53 23. Tsao MN, Rades D, Wirth A, Lo SS, Danielson BL, Vichare A, Hahn C, Chang EL (2012) International practice survey on the management of brain metastases: Third International Consensus Workshop on Palliative Radiotherapy and Symptom Control. Clin Oncol (R Coll Radiol) 24:e81–e92 24. Fife KM, Colman MH, Stevens GN, Firth IC, Moon D, Shannon KF, Harman R, Petersen-Schaefer K, Zacest AC, Besser M, Milton GW, McCarthy WH, Thompson JF (2004) Determinants of outcome in melanoma patients with cerebral metastases. J Clin Oncol 22:1293–1300