Abstract
Background
Prognostic estimates for patients with brain metastases (BM) stem from younger, healthier patients enrolled in clinical trials or databases from academic centers. We characterized population-level prognosis in elderly patients with BM.
Methods
Using Surveillance, Epidemiology, and End Results (SEER)–Medicare data, we identified 9882 patients ≥65 years old with BM secondary to lung, breast, skin, kidney, esophageal, colorectal, and ovarian primaries between 2014 and 2016. Survival was assessed by primary site and evaluated with Cox regression.
Results
In total, 2765 versus 7117 patients were diagnosed with BM at primary cancer diagnosis (synchronous BM, median survival = 2.9 mo) versus thereafter (metachronous BM, median survival = 3.4 mo), respectively. Median survival for all primary sites was ≤4 months, except ovarian cancer (7.5 mo). Patients with non-small-cell lung cancer (NSCLC) receiving epidermal growth factor receptor (EGFR)– or anaplastic lymphoma kinase (ALK)–based therapy for synchronous BM displayed notably better median survival at 12.5 and 20.1 months, respectively, versus 2.8 months exhibited by other patients with NSCLC; survival estimates in melanoma patients based on receipt of BRAF/MEK therapy versus not were 6.7 and 2.8 months, respectively. On multivariable regression, older age, greater comorbidity, and type of managing hospital were associated with poorer survival; female sex, higher median household income, and use of brain-directed stereotactic radiation, neurosurgical resection, or systemic therapy (versus brain-directed non-stereotactic radiation) were associated with improved survival (all P < 0.05).
Conclusions
Elderly patients with BM have a poorer prognosis than suggested by prior algorithms. If prognosis is driven by systemic and not intracranial disease, brain-directed therapy with potential for significant toxicity should be utilized cautiously.
Keywords: brain metastases, Medicare, population, prognosis, survival
Key Points.
Population-based survival estimates for patients with brain metastases are lacking.
Elderly patients with brain metastases have a median survival of only 3.2 months.
Aggressive, brain-directed treatments should be used cautiously among such patients.
Importance of the Study.
Current prognostic estimates for patients with BM stem from younger, healthier patients enrolled in clinical trials or managed at academic centers and may not generalize to the larger BM population. To our knowledge, ours is the first study to utilize SEER-Medicare data to characterize population-level survival outcomes associated with BM that develop either at or after primary cancer diagnosis. Among 9882 elderly patients with BM, we found median survival to be 3.2 months. Except for ovarian cancer, median survival was <4 months across all primary tumor sites. Only patients with NSCLC receiving EGFR- or ALK-based therapy for synchronous BM displayed notably better median survival at 12.5 and 20.1 months, respectively. Our work demonstrates that the prognosis for elderly patients with BM is poorer than previously reported. If driven by systemic and not intracranial disease progression, utilization of aggressive, brain-directed treatments should be reconsidered in such patients.
Brain metastases (BM) manifest in 20–40% of patients with solid malignancies1,2 and are generally associated with a guarded prognosis, varying based on primary site, patient age, performance status, viable treatment options, as well as extent of intracranial and extracranial disease.3–5 Prior investigations of prognosis among patients with BM have led to development of prognostic indices, including recursive partitioning analysis (RPA)3 and, more recently, the diagnosis specific–graded prognostic assessment (DS-GPA).6 While these classification systems have subsequently been validated,7–9 data from both the original and validation studies have largely stemmed from patients enrolled in Radiation Therapy Oncology Group (RTOG) clinical trials3,6 or institutional series from larger, academic centers.8–12 The generalizability of RPA and GPA-based indices to the larger population of patients with BM remains unclear. Such data may also be less applicable to elderly patients given their reduced ability to tolerate aggressive, brain-directed interventions.
The Surveillance, Epidemiology, and End Results (SEER) database recently released information regarding the presence versus absence of BM at diagnosis of primary malignancy, facilitating investigation of BM at a population-based level.13 However, given that most patients develop BM later in their disease course,14 such studies harbor limited ability to provide prognostic information for the majority of the BM population. Hence, better prognostication for such patients is needed, particularly among elderly patients, who are vulnerable to both disease and treatment-related sequelae of BM.15,16 Accordingly, we used SEER-Medicare data to characterize prognosis in elderly patients with BM on a population level.
Materials and Methods
Patient Population and Study Design
The SEER program captures data from 34.6% of patients with cancer in the United States and provides information on demographic, disease, and treatment-related covariates at diagnosis of primary malignancy.17 The SEER-Medicare database links SEER data to Medicare claims, facilitating investigations of disease progression and treatment beyond initial diagnosis.18
We used SEER-Medicare data to identify patients with BM secondary to solid malignancies that most commonly metastasize to the brain (i.e., lung, breast, melanoma, kidney, esophageal, colorectal, and ovarian cancers) between 2014 and 2016, the most recent year for which SEER-Medicare data are available. For patients with multiple primary tumors, the first primary was selected for evaluation. Patients with ≥3 claims associated with an ICD-9/ICD-10 diagnosis code for secondary malignant neoplasm of the brain, cerebral meninges, and spinal cord (ICD9-CM 198.3; ICD10-CM C79.31–79.32, N = 21 304) were ascribed to have BM, an approach associated with 97% sensitivity and 99% specificity for identifying patients with BM relative to manual chart review.19 The date of the first BM-associated ICD-9/ICD-10 code was designated as the date of diagnosis of BM, an approach shown to be 87% and 92% sensitive for predicting the actual date of BM diagnosis to within 15 and 30 days, respectively, compared with manual chart review.20
We excluded patients: (i) without continuous enrollment in Medicare Parts A and B in the year before BM diagnosis, (ii) with enrollment in a health maintenance organization in the year preceding BM diagnosis, (iii) whose date of death did not match exactly when comparing SEER and Medicare data, (iv) diagnosed with cancer at autopsy or death certificate, (v) <65 years at time of primary cancer diagnosis, (vi) for whom the date of last contact was not after the date of primary cancer diagnosis, and (vii) for whom the date of BM diagnosis was >15 days before the date of primary cancer diagnosis, leaving 9882 patients for analysis (Supplementary Figure 1).
To determine the initial, local treatment strategy for BM, we searched for relevant ICD-9/10 procedure, diagnosis, Current Procedural Terminology 4, and/or Healthcare Common Procedure Coding System (HCPCS) codes for neurosurgery, brain-directed stereotactic radiation, and brain-directed non-stereotactic radiation (Supplementary Table 1). We also examined the number of systemic agents received during a patient’s clinical course in a site-specific manner. Oral agents were assessed using Part D claims; intravenous agents were identified in National Claims History and outpatient files with drug-specific HCPCS codes.
We identified the number of distinct radiation courses to extracranial metastatic sites prior to BM diagnosis. Based on prior work,21 we defined a unique radiation therapy course via the presence of general radiotherapy (RT)-related codes (Supplementary Table 2) linked to a diagnosis code for metastatic disease at an extracranial site >14 days from the last RT-related code.
Statistical Methodology
We subsetted patients into those diagnosed with BM at (i.e., synchronous BM, N = 2765) versus after (i.e., metachronous BM, N = 7117) primary cancer diagnosis, and further stratified by primary cancer type. Patients with lung primaries were additionally divided into non-small-cell lung cancer (NSCLC) and small-cell lung cancer (SCLC). Categorical demographic, disease, and treatment-related characteristics were compared by primary tumor site using the chi-squared test. Normally and nonnormally distributed continuous covariates were compared across groups using ANOVA and the Kruskal–Wallis tests, respectively. Separate tables were generated for patients with synchronous versus metachronous BM. Median survival for each group was calculated with the Kaplan–Meier method; comparisons were made with the log-rank test.
For delineation of survival estimates, we further divided patients with NSCLC by histology and via the likely presence of targetable mutations in epidermal growth factor receptor (EGFR) or rearrangements in anaplastic lymphoma kinase (ALK).22 Since SEER does not contain a variable that directly provides information on EGFR or ALK status, we used receipt of any EGFR- or ALK-specific targeted therapy based on Part D claims as a surrogate. These therapies included erlotinib, gefitinib, afatinib, and osimertinib for EGFR mutations and crizotinib, alectinib, and ceritinib for ALK rearrangements. No patients received brigatinib or lorlatinib.
Similarly, we divided melanoma patients by BRAF status into those with a wild-type or unknown BRAF status and those with a BRAF mutant based on receipt of BRAF inhibitors, including vemurafenib, dabrafenib, trametinib, binimetinib, encorafenib, or cobimetinib. It is important to note, however, that some patients may have received an agent such as trametinib or binimetinib for an NRAS mutation,23 although we expect such instances to be rare given the era of this study. We also divided breast cancer patients into: human growth factor receptor 2 (HER2)–positive, hormone receptor–positive/HER2-negative, and triple-negative. Survival was calculated from date of BM diagnosis to date of death or censoring.
Predictors of all-cause mortality were assessed with univariable and multivariable Cox regression models adjusted for sex, race, marital status, high school completion rate (zip-code level), median household income (zip-code level), residence type (non-urban/unknown versus urban), Charlson comorbidity index (CCI, Deyo),24 primary tumor site, type of managing hospital (government, medical school-associated, and/or urban), initial BM treatment strategy, and for metachronous BM only, number of prior radiation regimens to sites of extracranial metastases. For covariates with significant violations of the proportional hazards assumption, the interaction of the covariate with time was included in the model. A 2-sided P-value <0.05 was considered statistically significant. This study was approved by our institutional review board. Analyses were performed using SAS version 9.4.
Results
Baseline Characteristics
Baseline characteristics by primary tumor site for patients with synchronous and metachronous BM are presented in Tables 1 and 2, respectively. There were significant differences across primary tumor sites for patients with synchronous BM with respect to age at diagnosis of primary cancer, age at diagnosis of BM, sex, race, marital status, high school completion rate and household income at the zip-code level, and type of managing hospital (government vs. not; medical school–associated vs. not; urban vs. not) (all P < 0.05). There were no significant differences in CCI (P > 0.05). For patients with metachronous BM, significant differences for all above variables by primary tumor site were noted (all P < 0.05).
Table 1.
Baseline characteristics of patients with brain metastases present at the time of primary cancer diagnosis
| NSCLCa (N = 2007) | SCLC (N = 444) | Breast (N = 71) | Melanoma (N = 109) | Kidney (N = 71) | Colorectal (N = 29) | Esophagus (N = 26) | P-value | |
|---|---|---|---|---|---|---|---|---|
| Age at diagnosis of primary malignancy, years, N (%) | 0.005 | |||||||
| <68 | 306 (15) | 73 (16) | 18 (25) | 16 (15) | 16 (23) | 29 (100) | 26 (100) | |
| 68–70 | 363 (18) | 100 (23) | 14 (20) | 12 (11) | 16 (23) | |||
| 71–74 | 461 (23) | 111 (25) | 13 (18) | 22 (20) | 14 (20) | |||
| 75–78 | 372 (19) | 77 (17) | 14 (20) | 20 (18) | 25 (35) | |||
| >78 | 505 (25) | 83 (19) | 12 (17) | 39 (36) | ||||
| Age at diagnosis of brain metastases, years, N (%) | 0.002 | |||||||
| <70 | 482 (24) | 119 (27) | 22 (31) | 20 (18) | 23 (32) | 12 (41) | 11 (42) | |
| 70–72 | 345 (17) | 99 (22) | 13 (18) | 18 (17) | 12 (17) | 17 (59) | 15 (58) | |
| 73–76 | 444 (22) | 96 (22) | 13 (18) | 19 (17) | 16 (23) | |||
| 77–81 | 383 (19) | 83 (19) | 23 (33) | 25 (23) | 20 (28) | |||
| >81 | 353 (18) | 47 (11) | 27 (25) | |||||
| Sex, N (%) | <0.001 | |||||||
| Male | 989 (49) | 228 (51) | 71 (100) | 74 (68) | 43 (61) | 17 (59) | 26 (100) | |
| Female | 1018 (51) | 216 (49) | 35 (32) | 28 (39) | 12 (41) | |||
| Race, N (%) | <0.001 | |||||||
| White | 1602 (80) | 380 (86) | 53 (75) | 109 (100) | 56 (79) | 29 (100) | 26 (100) | |
| African American | 185 (9) | 36 (8) | 18 (25) | 15 (21) | ||||
| Hispanic | 94 (5) | 28 (6) | ||||||
| Asian/Pacific Islander | 126 (6) | |||||||
| Other/Unknown | ||||||||
| Marital status, N (%) | 0.003 | |||||||
| Married/domestic partnership | 1009 (50) | 218 (49) | 23 (32) | 64 (59) | 36 (51) | 11 (38) | 26 (100) | |
| Unmarried/single | 942 (47) | 226 (51) | 48 (68) | 45 (42) | 35 (49) | 18 (62) | ||
| Unknown | 56 (3) | |||||||
| Type of residence, N (%) | 0.95 | |||||||
| Urban | 1756 (87) | 384 (86) | 71 (100) | 97 (89) | 71 (100) | 29 (100) | 26 (100) | |
| Non-urban | 251 (13) | 60 (14) | 12 (11) | |||||
| Unknown | ||||||||
| Graduated from high school, median % (IQR) b | 86 (78–92) | 85 (76–91) | 85 (77–92) | 89 (82–94) | 89 (81–93) | 83 (77–91) | 89 (80–92) | 0.004 |
| Household income (per 10K USD), median (IQR) b | 4.8 (3.6–6.5) | 4.6 (3.5–6.3) | 4.7 (3.6–6.9) | 5.5 (4.0–7.5) | 5.6 (4.3–7.7) | 3.9 (3.1–6.0) | 5.3 (4.5–6.6) | 0.003 |
| Charlson comorbidity index, N (%) c | 0.34 | |||||||
| 0–2 | 1599 (80) | 348 (78) | 71 (100) | 90 (83) | 57 (80) | 29 (100) | 26 (100) | |
| >2 | 337 (17) | 96 (22) | 19 (17) | 14 (20) | ||||
| Unknown | 71 (4) | |||||||
| Government hospital, N (%) | 0.01 | |||||||
| No | 1356 (68) | 309 (70) | 58 (82) | 80 (73) | 42 (59) | 29 (100) | 26 (100) | |
| Yes | 500 (25) | 113 (25) | 13 (18) | 29 (27) | 29 (41) | |||
| Unknown | 151 (8) | 22 (5) | ||||||
| Medical school‒associated hospital, N (%) | 0.03 | |||||||
| No | 635 (32) | 158 (36) | 25 (35) | 30 (28) | 17 (24) | 14 (48) | 26 (100) | |
| Yes | 1221 (61) | 264 (59) | 46 (65) | 79 (72) | 42 (59) | 15 (52) | ||
| Unknown | 151 (8) | 22 (5) | 12 (17) | |||||
| Urban hospital, N (%) | 0.01 | |||||||
| No | 221 (11) | 60 (14) | 20 (28) | 14 (13) | 17 (24) | 29 (100) | 26 (100) | |
| Unknown | 149 (7) | 22 (5) | ||||||
| Yes | 1637 (82) | 362 (82) | 51 (72) | 95 (87) | 54 (76) |
Abbreviations: IQR, interquartile range; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
aIncludes lung primaries that are not specifically listed as adenocarcinoma, squamous cell, adenosquamous, large cell, or bronchoalveolar (a histology still designated by SEER).
bZip-code level.
cExcluded diagnosis of metastatic cancer so as not to inflate all scores by 6 points.
Notes: (1) Subgroups for age-related variables were based upon division of the entire patient population into quintiles. (2) Categories for certain variables were grouped together so as to comply with NCI data policy of not displaying any cells with values <11. For the same reason, we could not display baseline characteristics for patients with ovarian cancer. (3) Percentages may not add up to 100 due to rounding.
Table 2.
Baseline characteristics of patients with brain metastases identified after primary cancer diagnosis
| NSCLCa (N = 2994) | SCLC (N = 839) | Breast (N = 1294) | Melanoma (N = 811) | Kidney (N = 351) | Colorectal (N = 560) | Esophagus (N = 157) | P-value | |
|---|---|---|---|---|---|---|---|---|
| Age at diagnosis of primary malignancy, years, N (%) | <0.001 | |||||||
| <68 | 567 (19) | 190 (23) | 345 (27) | 150 (19) | 84 (24) | 115 (21) | 28 (18) | |
| 68–70 | 602 (20) | 185 (22) | 276 (21) | 143 (18) | 74 (21) | 111 (20) | 40 (25) | |
| 71–74 | 695 (23) | 211 (25) | 291 (22) | 174 (21) | 70 (20) | 110 (20) | 52 (33) | |
| 75–78 | 521 (17) | 153 (18) | 177 (14) | 129 (16) | 66 (19) | 111 (20) | 19 (12) | |
| >78 | 609 (20) | 100 (12) | 205 (16) | 215 (27) | 57 (16) | 113 (20) | 18 (11) | |
| Age at diagnosis of brain metastases, years, N (%) | <0.001 | |||||||
| <70 | 525 (18) | 222 (26) | 141 (11) | 64 (8) | 60 (17) | 49 (9) | 26 (17) | |
| 70–72 | 555 (19) | 182 (22) | 215 (17) | 103 (13) | 63 (18) | 73 (13) | 38 (24) | |
| 73–76 | 712 (24) | 208 (25) | 325 (25) | 173 (21) | 70 (20) | 120 (21) | 46 (29) | |
| 77–81 | 624 (21) | 154 (18) | 306 (24) | 192 (24) | 85 (24) | 157 (28) | 27 (17) | |
| >81 | 578 (19) | 73 (9) | 307 (24) | 279 (34) | 73 (21) | 161 (29) | 20 (13) | |
| Sex, N (%) | <0.001 | |||||||
| Male | 1423 (48) | 368 (44) | 19 (1) | 582 (72) | 244 (70) | 317 (57) | 122 (78) | |
| Female | 1571 (52) | 471 (56) | 1275 (99) | 229 (28) | 107 (30) | 243 (43) | 35 (22) | |
| Race, N (%) | <0.001 | |||||||
| White | 2392 (80) | 739 (88) | 1054 (81) | 782 (96) | 288 (82) | 446 (80) | 136 (87) | |
| African American | 249 (8) | 46 (5) | 122 (9) | 17 (2) | 20 (6) | 44 (8) | 21 (13) | |
| Hispanic | 118 (4) | 33 (4) | 66 (5) | 26 (7) | 39 (7) | |||
| Asian/Pacific Islander | 218 (7) | 21 (3) | 52 (4) | 12 (1) | 17 (5) | 31 (6) | ||
| Other/Unknown | 17 (1) | |||||||
| Marital status, N (%) | <0.001 | |||||||
| Married/domestic partnership | 1632 (55) | 427 (51) | 614 (47) | 471 (58) | 223 (64) | 312 (56) | 90 (57) | |
| Unmarried/single | 1215 (41) | 379 (45) | 617 (48) | 205 (25) | 111 (32) | 218 (39) | 67 (43) | |
| Unknown | 147 (5) | 33 (4) | 63 (5) | 135 (17) | 17 (5) | 30 (5) | ||
| Type of residence, N (%) | 0.003 | |||||||
| Urban | 2696 (90) | 722 (86) | 1179 (91) | 737 (91) | 308 (88) | 494 (88) | 140 (89) | |
| Non-urban | 298 (10) | 117 (14) | 115 (9) | 74 (9) | 43 (12) | 66 (12) | 17 (11) | |
| Unknown | ||||||||
| Graduated from high school, median % (IQR) b | 86 (78–92) | 85 (76–91) | 88 (79–92) | 90 (82–94) | 86 (78–92) | 86 (77–92) | 88 (81–92) | <0.001 |
| Household income (per 10K USD), median (IQR) b | 4.9 (3.7–6.8) | 4.7 (3.4–6.6) | 5.1 (3.8–7.0) | 5.7 (4.1–7.5) | 5.0 (3.6–7.0) | 4.9 (3.5–6.8) | 5.4 (3.8–7.5) | <0.001 |
| Charlson comorbidity index, N (%) c | <0.001 | |||||||
| 0–2 | 2165 (72) | 573 (68) | 1052 (81) | 636 (78) | 216 (62) | 443 (79) | 125 (80) | |
| >2 | 808 (27) | 266 (32) | 242 (19) | 175 (22) | 135 (38) | 117 (21) | 32 (20) | |
| Unknown | 21 (1) | |||||||
| Government hospital, N (%) | <0.001 | |||||||
| No | 2172 (73) | 620 (74) | 944 (73) | 536 (66) | 237 (68) | 396 (71) | 118 (75) | |
| Yes | 689 (23) | 193 (23) | 311 (24) | 232 (29) | 100 (28) | 129 (23) | 39 (25) | |
| Unknown | 133 (4) | 26 (3) | 39 (3) | 43 (5) | 14 (4) | 35 (6) | ||
| Medical school‒associated hospital, N (%) | <0.001 | |||||||
| No | 870 (29) | 268 (32) | 376 (29) | 179 (22) | 87 (25) | 158 (28) | 40 (25) | |
| Yes | 1991 (67) | 545 (65) | 880 (68) | 590 (73) | 250 (71) | 368 (66) | 117 (75) | |
| Unknown | 133 (4) | 26 (3) | 38 (3) | 42 (5) | 14 (4) | 34 (6) | ||
| Urban hospital, no. (%) | 0.04 | |||||||
| No | 288 (10) | 83 (10) | 125 (10) | 61 (8) | 29 (8) | 67 (12) | 12 (8) | |
| Yes | 2573 (86) | 730 (87) | 1131 (87) | 708 (87) | 308 (88) | 459 (82) | 145 (92) | |
| Unknown | 133 (4) | 26 (3) | 38 (3) | 42 (5) | 14 (4) | 34 (6) |
Abbreviations: IQR, interquartile range; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
aIncludes lung primaries that are not specifically listed as adenocarcinoma, squamous cell, adenosquamous, large cell, or bronchoalveolar (a histology still designated by SEER).
bZip-code level.
cExcluded diagnosis of metastatic cancer so as not to inflate all scores by 6 points.
Notes: (1) Subgroups for age-related variables were based upon division of the entire patient population into quintiles. (2) Categories for certain variables were grouped together so as to comply with NCI data policy of not displaying any cells with values <11. For the same reason, we could not display baseline characteristics for patients with ovarian cancer. (3) Percentages may not add up to 100 due to rounding.
Extent of Disease and BM Management Strategies
For patients with synchronous BM, the majority also had metastases in the liver, lung, or bones at time of BM diagnosis (n = 1460 patients; 52.8%; Supplementary Table 3). For patients with metachronous BM, most had received multiple systemic agents prior to BM diagnosis (Supplementary Table 4). The most commonly utilized initial treatment strategy (Supplementary Tables 3, 4) for BM among patients with synchronous or metachronous BM was brain-directed non-stereotactic radiation inclusive of whole brain radiation; rates of such therapy by primary tumor site ranged from 44% to 58%.
Survival Estimates
The median survival for the entire cohort of patients from BM diagnosis was 3.2 months, with a 1-year survival rate of 21.8% (95% CI, 21.0–22.6%). Among patients with synchronous versus metachronous BM, median survival was 2.9 versus 3.4 months, respectively, and 1-year survival rates were 18.4% (17.0–19.8%) versus 23.2% (22.2–24.2%), respectively. Except for patients with ovarian cancer, who had a median survival of 7.5 months, median survival across all primary cancer sites was 4 months or less. One-year survival rates by primary site were: NSCLC, 21.2% (20.1–22.4%); SCLC, 17.0% (15.0–19.1%); breast, 31.1% (28.6–33.5%); melanoma, 20.9% (18.3–23.6%); kidney, 23.0% (19.1–27.1%); colorectal, 16.5% (13.6–19.6%); esophageal, 12.6% (8.3–17.8%); and ovarian, 37.0% (28.4–45.6%). Survival estimates of patients with synchronous and metachronous BM by primary cancer site are presented in Table 3 and Supplementary Figure 2.
Table 3.
Survival outcomes of patients with brain metastases by primary cancer site
| Brain Metastases At Time Of Primary Cancer Diagnosis | Brain Metastases After Primary Cancer Diagnosis | |||
|---|---|---|---|---|
| Primary Tumor Site | Median Survival in Months from Time of BM Diagnosis (95% CI) | 1-Year Survival from Time of BM Diagnosis | Median Survival In Months from Time of BM Diagnosis (95% CI) | 1-Year Survival from Time of BM Diagnosis |
| Lung | 2.9 (2.7–3.1) | 18.2% | 3.3 (3.1–3.5) | 21.7% |
| Adenocarcinoma | 3.8 (3.3–4.1) | 25.0% | 3.7 (3.4–4.0) | 25.3% |
| Squamous cell carcinoma | 2.2 (1.9–2.9) | 11.4% | 2.8 (2.6–3.4) | 17.0% |
| Small cell carcinoma | 3.0 (2.6–3.7) | 12.6% | 3.6 (3.2–4.0) | 19.3% |
| Other NSCLC/ unspecified | 1.9 (1.7–2.2) | 12.3% | 2.7 (2.4–3.1) | 20.7% |
| ALK rearrangeda | 20.1 (1.3-NR) | 55.6% | 6.0 (3.7–16.2) | 39.4% |
| EGFR mutanta | 12.5 (7.5–16.3) | 51.9% | 4.2 (3.4–5.5) | 29.5% |
| All othersb | 2.8 (2.5–3.1) | 17.7% | 3.3 (3.0–3.6) | 21.9% |
| Breast | 2.1 (1.6–3.5) | 22.5% | 4.5 (3.9–4.9) | 31.5% |
| HER2-positive | 2.5 (1.3–9.4) | 23.8% | 6.4 (3.8–7.7) | 36.3% |
| HR-positive/HER2- negative | 2.0 (1.1–9.6) | 29.0% | 4.9 (4.1–6.8) | 34.5% |
| Triple-negative | 2.3 (0.1–10.1) | 11.1% | 3.4 (2.2–4.7) | 21.4% |
| Unknown | 1.3 (0.2–2.2) | 10.0% | 4.0 (3.5–4.9) | 30.5% |
| Melanoma | 3.0 (2.4–3.9) | 20.2% | 2.8 (2.3–3.1) | 21.0% |
| BRAF mutantc,d | 6.7 (2.1–13.7) | 28.6% | 2.8 (2.3–3.7) | 14.3% |
| BRAF wild-type or unknowne | 2.8 (1.8–4.7) | 21.0% | 2.9 (2.3–3.3) | 22.7% |
| Kidney | 1.8 (1.3–2.3) | 12.7% | 3.5 (2.7–4.4) | 25.1% |
| Colorectal | 3.0 (1.4–7.2) | 20.7% | 2.5 (2.3–3.0) | 16.3% |
| Esophagus | 4.0 (2.4–6.6) | 19.2% | 2.3 (1.9–3.2) | 11.5% |
| Ovarian | 7.7 (1.5–31.7) | 37.5% | 7.5 (4.4–9.8) | 36.9% |
Abbreviations: ALK, anaplastic lymphoma kinase; BM, brain metastases; CI, confidence interval; EGFR, epidermal growth factor receptor; HER2, human epidermal growth factor receptor 2; HR, hormone receptor; MEK, mitogen-activated protein kinase; NR, not reached; NSCLC, non-small cell lung cancer.
aBased on patients with Part D claims (n = 7195) with NSCLC as their primary tumor (n = 3628) and who received relevant targeted agents per Part D file at any point from date of primary cancer diagnosis up to 60 days after BM diagnosis (EGFR = 450; ALK = 42).
bComparison group includes those NSCLC patients with Part D claims but without evidence of receipt of EGFR or ALK-targeting therapies (n = 3136).
cBased on patients with Part D claims (n = 7195) and melanoma as their primary tumor (n = 631) who received BRAF/MEK-targeting agents per Part D file at any point from date of primary cancer diagnosis up to 60 days after BM diagnosis (n = 63).
dMay include some patients with NRAS mutations who received MEK inhibitors as monotherapy, although we suspect that this is a small percentage given that relatively few patients have such mutations and validating clinical studies were largely published after the timeframe studied here.
eComparison group includes those melanoma patients with Part D claims but without evidence of receipt of BRAF/MEK-targeting therapies (n = 568).
Among patients with synchronous BM secondary to NSCLC, there were significant differences in survival between those receiving ALK-based therapy, EGFR-based therapy, or neither (20.1, 12.5, and 2.8 mo, respectively, P < 0.001; Supplementary Figure 3). The same trend was notable, but to a lesser degree, for patients with NSCLC-associated metachronous BM. Similarly, among melanoma patients with synchronous BM and a BRAF mutation, median survival was 6.7 months compared with 2.8 months in patients who did not receive a BRAF/MEK agent; median survival for melanoma patients with metachronous BM and mutant BRAF was considerably lower than mutant patients with synchronous BM (Supplementary Figure 4).
Predictors of Overall Survival
Synchronous brain metastases
Cox regression analyses for all-cause mortality of patients with synchronous BM are displayed in Table 4. In the adjusted model, age >81 or age of 77–81 versus age <70 (HR: 1.57 [95% CI, 1.39–1.78], P < 0.001; HR: 1.18 [1.05–1.33], P = 0.008, respectively), CCI > 2 versus 0–2 (HR: 1.20 [1.08–1.34], P < 0.001), kidney as the primary cancer site versus the reference of NSCLC (HR: 1.41 [1.10–1.81], P = 0.007), type of managing hospital (non-government/not medical school‒associated versus the reference of government and medical school-associated; HR: 1.18 [1.03–1.34], P = 0.01), and lack of treatment for BM (HR: 1.45 [1.31–1.60], P < 0.001) were associated with worse survival. Female sex (HR: 0.89 [0.82–0.97], P = 0.005), African American or Asian race versus the reference of White race (HR: 0.86 [0.74–0.99], P = 0.04; HR 0.60 [0.49–0.72], P < 0.001, respectively), marital status (married/partnered versus unmarried/single; HR: 0.89 [0.82–0.97], P = 0.005), a higher zip-code level household income (HR: 0.97 [0.95–0.99], P = 0.001), receiving care at an urban versus non-urban hospital (HR: 0.86 [0.75–0.97], P = 0.02), and receipt of brain-directed stereotactic radiation (HR: 0.52 [0.46–0.59], P < 0.001), neurosurgical resection (HR: 0.56 [0.49–0.64], P < 0.001), or systemic therapy without local brain-directed therapy (HR: 0.79 (0.68–0.93), P = 0.004) versus the reference of brain-directed non-stereotactic RT as initial BM treatment strategy were associated with longer survival.
Table 4.
Univariable and multivariable Cox regression for all-cause mortality of patients with brain metastases present at the time of primary cancer diagnosis
| Univariable HR (95% CI) | P value | Multivariable HR (95% CI) | P value | |
|---|---|---|---|---|
| Age at BM diagnosis, years | ||||
| <70 | Ref | Ref | ||
| 70–72 | 1.01 (0.90–1.14) | 0.84 | 1.02 (0.90–1.15) | 0.81 |
| 73–76 | 1.09 (0.97–1.22) | 0.15 | 1.12 (0.99–1.25) | 0.06 |
| 77–81 | 1.16 (1.03–1.31) | 0.01 | 1.18 (1.05–1.33) | 0.008 |
| >81 | 1.70 (1.50–1.91) | <0.001 | 1.57 (1.39–1.78) | <0.001 |
| Sex | 0.02 | 0.005 | ||
| Male | Ref | Ref | ||
| Female | 0.91 (0.85–0.99) | 0.89 (0.82–0.97) | ||
| Race | ||||
| White | Ref | Ref | ||
| African American | 1.01 (0.88–1.16) | 0.91 | 0.86 (0.74–0.99) | 0.04 |
| Hispanic | 1.08 (0.91–1.29) | 0.39 | 0.96 (0.80–1.16) | 0.68 |
| Asian | 0.59 (0.49–0.72) | <0.001 | 0.60 (0.49–0.72) | <0.001 |
| Other/Unknown | 0.94 (0.54–1.61) | 0.81 | 1.08 (0.62–1.87) | 0.79 |
| Marital status at diagnosis | ||||
| Unmarried/single | Ref | Ref | ||
| Married/partnered | 0.81 (0.75–0.87) | <0.001 | 0.89 (0.82–0.97) | 0.005 |
| Unknown | 0.93 (0.74–1.16) | 0.51 | 1.06 (0.84–1.33) | 0.63 |
| Graduated from high school (per % increase) a | 0.99 (0.99–0.99) | <0.001 | 1.00 (0.99–1.00) | 0.55 |
| Household income (per 10K USD increase) a | 0.95 (0.94–0.97) | <0.001 | 0.97 (0.95–0.99) | 0.001 |
| Residence | <0.001 | 0.13 | ||
| Non-urban/ Unknown | Ref | Ref | ||
| Urban | 0.77 (0.69–0.87) | 0.90 (0.79–1.03) | ||
| Charlson comorbidity index b | ||||
| 0–2 | Ref | Ref | ||
| >2 | 1.33 (1.20–1.47) | <0.001 | 1.20 (1.08–1.34) | <0.001 |
| Unknown | 1.10 (0.89–1.36) | 0.36 | 0.97 (0.78–1.21) | 0.80 |
| Primary tumor site | ||||
| NSCLC | Ref | Ref | ||
| SCLC | 1.15 (1.03–1.27) | 0.01 | 1.03 (0.92–1.14) | 0.65 |
| Breast | 0.93 (0.73–1.19) | 0.56 | 0.98 (0.76–1.26) | 0.85 |
| Melanoma | 0.86 (0.70–1.06) | 0.16 | 0.96 (0.78–1.19) | 0.72 |
| Kidney | 1.35 (1.06–1.73) | 0.02 | 1.41 (1.10–1.81) | 0.007 |
| Colorectal | 0.91 (0.62–1.34) | 0.64 | 0.97 (0.66–1.44) | 0.89 |
| Esophagus | 0.57 (0.30–1.09) | 0.09 | 0.63 (0.32–1.22) | 0.17 |
| Esophagus*log(time)c | 1.60 (1.11–2.31) | 0.01 | 1.65 (1.12–2.41) | 0.01 |
| Ovarian | 0.57 (0.27–1.19) | 0.13 | 0.56 (0.27–1.19) | 0.13 |
| Type of managing hospital d | ||||
| Government and medical school‒associated | Ref | Ref | ||
| Government and not medical school‒associated | 1.25 (1.06–1.47) | 0.007 | 0.99 (0.84–1.17) | 0.87 |
| Non-government and medical school‒associated | 0.98 (0.87–1.09) | 0.70 | 1.00 (0.89–1.12) | 0.97 |
| Non-government and not medical school‒associated | 1.30 (1.14–1.47) | <0.001 | 1.18 (1.03–1.34) | 0.01 |
| Unknown | 2.85 (2.40–3.39) | <0.001 | 18.3 (4.50–74.20) | <0.001 |
| Urban hospital | ||||
| No | Ref | Ref | ||
| Yes | 0.91 (0.81–1.02) | 0.12 | 0.86 (0.75–0.97) | 0.02 |
| Unknown | 0.28 (0.24–0.33) | <0.001 | 0.10 (0.03–0.42) | 0.002 |
| Initial BM treatment strategy e | ||||
| Non-stereotactic brain-directed radiation | Ref | Ref | ||
| Stereotactic brain-directed radiation | 0.50 (0.44–0.57) | <0.001 | 0.52 (0.46–0.59) | <0.001 |
| Neurosurgical resection | 0.57 (0.51–0.65) | <0.001 | 0.56 (0.49–0.64) | <0.001 |
| Systemic therapy without local brain-directed therapyf | 0.75 (0.65–0.88) | <0.001 | 0.79 (0.68–0.93) | 0.004 |
| Other/noneg | 1.67 (1.52–1.84) | <0.001 | 1.45 (1.31–1.60) | <0.001 |
Abbreviations: BM, brain metastasis; CI, confidence interval; HR, hazard ratio; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
aZip-code level.
bExcluded diagnosis of metastatic cancer so as not to inflate all scores by 6 points.
cFor covariates with significant violations of the proportional hazards assumption, the interaction of the covariate with time (denoted by *) was included in the model.
dCovariates for medical school-associated hospital and government hospital were re-grouped in this manner due to high linkage between ‘unknown’ categories for these variables.
ePatients were considered recipients of a treatment if at least one of the relevant administrative codes (Table A1) was present within 1 month prior—2 months after BM diagnosis for brain-directed radiation or 1 month prior—1 month after BM diagnosis for neurosurgical resection.
fMay underestimate patients receiving systemic therapy given that data on receipt of oral systemic agents were only available for patients with Part D coverage.
gMay include patients who received systemic therapy but for whom this information was not captured by claims due to lack of Part D coverage.
Metachronous brain metastases
Univariable and multivariable Cox regression analyses for all-cause mortality of patients with BM after primary cancer diagnosis are displayed in Table 5. In the adjusted model, age >81 versus age <70 (HR: 1.19 [1.09–1.30], P < 0.001), CCI > 2 versus 0–2 (HR: 1.25 [1.18–1.32], P < 0.001), melanoma or esophagus as the primary cancer versus NSCLC (HR: 1.13 [1.03–1.25], P = 0.01; HR: 1.41 [1.20–1.67], P < 0.001, respectively), and type of managing hospital (government/not medical school‒associated or non-government/not medical school-associated versus government and medical school‒associated) (HR: 1.32 [1.18–1.48], P < 0.001; HR: 1.24 [1.14–1.34], P < 0.001, respectively), and number of prior radiation therapy regimens to sites of extracranial metastases (one versus none; HR: 1.10 [1.01–1.19], P = 0.02) were associated with worse survival. Female sex (HR: 0.90 [0.84–0.95], P < 0.001), Asian race versus white race (HR: 0.82 [0.73–0.93], P = 0.001), a higher zip-code level household income (HR: 0.98 [0.97–0.99], P < 0.001), living in an urban residence versus a non-urban/unknown residence (HR: 0.87 [0.80–0.95], P = 0.003), SCLC (HR: 0.89 [0.82–0.97], P = 0.005), breast (HR: 0.80 [0.74–0.86], P < 0.001), kidney (HR: 0.88 [0.79–1.00], P = 0.04), or ovarian (HR: 0.76 [0.61–0.94], P = 0.01) as the primary cancer site versus NSCLC, and brain-directed stereotactic radiation (HR: 0.58 [0.54–0.63], P < 0.001), neurosurgery (HR: 0.54 [0.49–0.59], P < 0.001), or systemic therapy without local brain-directed therapy (HR: 0.74 [0.68–0.81], P < 0.001) versus brain-directed non-stereotactic RT as initial BM treatment strategies were associated with longer survival.
Table 5.
Univariable and multivariable Cox regression models for all-cause mortality of patients with brain metastases identified after primary cancer diagnosis
| Univariable HR (95% CI) | P-value | Multivariable OR (95% CI) | P-value | |
|---|---|---|---|---|
| Age at BM diagnosis, years | ||||
| <70 | Ref | Ref | ||
| 70–72 | 1.04 (0.95–1.13) | 0.41 | 1.03 (0.94–1.12) | 0.57 |
| 73–76 | 1.02 (0.94–1.11) | 0.62 | 1.02 (0.94–1.10) | 0.67 |
| 77–81 | 1.10 (1.02–1.20) | 0.02 | 1.08 (0.99–1.17) | 0.07 |
| >81 | 1.29 (1.19–1.40) | <0.001 | 1.19 (1.09–1.30) | <0.001 |
| Sex | <0.001 | <0.001 | ||
| Male | Ref | Ref | ||
| Female | 0.83 (0.79–0.87) | 0.90 (0.84–0.95) | ||
| Race | ||||
| White | Ref | Ref | ||
| African American | 1.07 (0.97–1.17) | 0.19 | 1.02 (0.93–1.13) | 0.67 |
| Hispanic | 0.93 (0.82–1.05) | 0.24 | 0.92 (0.81–1.05) | 0.21 |
| Asian | 0.82 (0.73–0.93) | 0.001 | 0.82 (0.73–0.93) | 0.001 |
| Other/unknown | 0.53 (0.36–0.79) | 0.002 | 0.49 (0.33–0.72) | <0.001 |
| Marital status at diagnosis | ||||
| Unmarried/single | Ref | Ref | ||
| Married/partnered | 0.94 (0.89–0.99) | 0.01 | 0.96 (0.90–1.01) | 0.09 |
| Unknown | 0.91 (0.82–1.01) | 0.08 | 0.92 (0.82–1.03) | 0.13 |
| Graduated from high school (per % increase) a | 0.99 (0.99–1.00) | <0.001 | 1.00 (1.00–1.00) | 0.96 |
| Household income (per 10K USD increase) a | 0.97 (0.96–0.98) | <0.001 | 0.98 (0.97–0.99) | <0.001 |
| Residence | <0.001 | 0.003 | ||
| Non-urban/unknown | Ref | Ref | ||
| Urban | 0.80 (0.74–0.87) | 0.87 (0.80–0.95) | ||
| Charlson comorbidity index b | ||||
| 0–2 | Ref | Ref | ||
| >2 | 1.32 (1.25–1.40) | <0.001 | 1.25 (1.18–1.32) | <0.001 |
| Unknown | 1.49 (1.10–2.02) | 0.01 | 0.99 (0.73–1.36) | 0.97 |
| Primary tumor site | ||||
| NSCLC | Ref | Ref | ||
| SCLC | 1.02 (0.94–1.11) | 0.63 | 0.89 (0.82–0.97) | 0.005 |
| Breast | 0.79 (0.74–0.85) | <0.001 | 0.80 (0.74–0.86) | <0.001 |
| Melanoma | 1.12 (1.02–1.23) | 0.02 | 1.13 (1.03–1.25) | 0.01 |
| Melanoma*log(time)c | 0.90 (0.85–0.95) | <0.001 | 0.92 (0.87–0.97) | 0.003 |
| Kidney | 0.93 (0.83–1.05) | 0.23 | 0.88 (0.79–1.00) | 0.04 |
| Colorectal | 1.13 (1.03–1.24) | 0.01 | 1.05 (0.96–1.16) | 0.30 |
| Esophagus | 1.35 (1.15–1.60) | <0.001 | 1.41 (1.20–1.67) | <0.001 |
| Ovarian | 0.66 (0.54–0.82) | <0.001 | 0.76 (0.61–0.94) | 0.01 |
| Type of managing hospital d | ||||
| Government and medical school‒associated | Ref | Ref | ||
| Government and not medical school‒associated | 1.40 (1.25–1.56) | <0.001 | 1.32 (1.18–1.48) | <0.001 |
| Non-government and medical school‒associated | 0.99 (0.93–1.06) | 0.79 | 1.04 (0.97–1.12) | 0.23 |
| Non-government and not medical school‒associated | 1.26 (1.17–1.37) | <0.001 | 1.24 (1.14–1.34) | <0.001 |
| Unknown | 2.33 (2.04–2.65) | <0.001 | 1.31 (0.42–4.10) | 0.65 |
| Urban hospital | ||||
| No | Ref | Ref | ||
| Yes | 0.89 (0.82–0.96) | 0.003 | 0.93 (0.85–1.01) | 0.09 |
| Unknown | 0.21 (0.19–0.23) | <0.001 | 1.52 (0.48–4.80) | 0.47 |
| RT regimens to ECM prior to BM diagnosis | ||||
| None | Ref | Ref | ||
| 1 | 1.06 (0.99–1.15) | 0.12 | 1.10 (1.01–1.19) | 0.02 |
| 2+ | 1.01 (0.89–1.14) | 0.93 | 1.02 (0.90–1.16) | 0.72 |
| Initial BM treatment strategy e | ||||
| Non-stereotactic brain-directed radiation | Ref | Ref | ||
| Stereotactic brain-directed radiation | 0.58 (0.53–0.63) | <0.001 | 0.58 (0.54–0.63) | <0.001 |
| Neurosurgical resection | 0.55 (0.50–0.60) | <0.001 | 0.54 (0.49–0.59) | <0.001 |
| Systemic therapy without local brain-directed therapyf | 0.72 (0.66–0.78) | <0.001 | 0.74 (0.68–0.81) | <0.001 |
| Other/noneg | 1.05 (0.99–1.11) | 0.13 | 1.00 (0.94–1.06) | 0.98 |
Abbreviations: BM, brain metastasis; CI, confidence interval; ECM, extracellular matrix; HR, hazard ratio; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
aZip-code level.
bExcluded diagnosis of metastatic cancer so as not to inflate all scores by 6 points.
cFor covariates with significant violations of the proportional hazards assumption, the interaction of the covariate with time (denoted by *) was included in the model.
dCovariates for medical school-associated hospital and government hospital were regrouped in this manner due to high linkage between ‘unknown’ categories for these variables.
ePatients were considered recipients of a treatment if at least one of the relevant administrative codes (Table A1) was present within 1 month prior—2 months after BM diagnosis for brain-directed radiation or 1 month prior—1 month after BM diagnosis for neurosurgical resection.
fMay underestimate patients receiving systemic therapy given that data on receipt of oral systemic agents were only available for patients with Part D coverage.
gMay include patients who received systemic therapy but for whom this information was not captured by claims due to lack of Part D coverage.
Survival by treatment strategy
The median survival and univariable Cox regression analyses for all-cause mortality by primary tumor site and initial BM management strategy are presented in Table 6. Among patients with synchronous BM, receipt of stereotactic brain-directed radiation was associated with improved survival compared with the reference of non-stereotactic brain-directed radiation among those with NSCLC (HR: 0.50 [0.43–0.57], P < 0.001), SCLC (HR: 0.44 [0.25–0.75], P = 0.003), and melanoma (HR: 0.51 [0.27–0.94], P = 0.03). Neurosurgical resection was associated with improved survival among those with NSCLC (HR: 0.58 [0.50–0.67], P < 0.001), melanoma (HR: 0.41 [0.24–0.70], P = 0.001), and esophageal (HR: 0.37 [0.14–0.99], P = 0.05) primary tumors. Among patients with synchronous BM, systemic therapy without local brain-directed therapy was associated with improved survival for patients with NSCLC (HR: 0.74 [0.61–0.91], P = 0.003). Lack of treatment as captured by claims data was associated with poorer survival for patients with synchronous BM and NSCLC (HR: 1.64 [1.46–1.83], P < 0.001), SCLC (HR: 2.05 [1.61–2.60], P < 0.001), or breast cancer (HR: 3.49 [1.82–6.67], P < 0.001).
Table 6.
Survival estimates and univariable regression models for all-cause mortality of patients with brain metastases by primary tumor type
| BM at time of primary cancer diagnosis | NSCLCa (N = 2007) | SCLC (N = 444) | Breast (N = 71) | Melanoma (N = 109) | Kidney (N = 71) | Colorectal (N = 29) | Esophagus (N = 26) | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Median survival (mo.) |
HR (95% CI) | P value | Median survival (mo.) |
HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | |
| Initial BM management strategy, N (%) b | |||||||||||||||||||||
| Any non-stereotactic brain-directed radiation | 2.5 | Ref | -- | 3.0 | Ref | -- | 2.3 | Ref | -- | 2.0 | Ref | -- | 2.3 | Ref | -- | 2.8 | Ref | -- | 2.7 | Ref | -- |
| Any stereotactic brain-directed radiation | 7.6 | 0.50 (0.43–0.57) | <0.001 | 8.8 | 0.44 (0.25–0.75) | 0.003 | 10.1 | 0.70 (0.34–1.46) | 0.34 | 5.3 | 0.51 (0.27–0.94) | 0.03 | 2.4 | 0.59 (0.30–1.19) | 0.14 | 11.3 | 0.58 (0.21–1.64) | 0.31 | 12.6 | 0.15 (0.02–1.35) | 0.09 |
| Any neurosurgical resection | 6.7 | 0.58 (0.50–0.67) | <0.001 | 4.6 | 0.75 (0.51–1.09) | 0.13 | 4.4 | 0.72 (0.30–1.75) | 0.47 | 8.7 | 0.41 (0.24–0.70) | 0.001 | 2.4 | 0.79 (0.35–1.78) | 0.57 | 8.4 | 0.44 (0.10–1.95) | 0.28 | 8.3 | 0.37 (0.14–0.99) | 0.05 |
| Systemic therapy without local brain-directed therapyc | 4.3 | 0.74 (0.61–0.91) | 0.003 | 5.7 | 0.83 (0.63–1.10) | 0.19 | 18.5 | 0.49 (0.19–1.28) |
0.15 | 4.4 | 0.81 (0.28–2.30) | 0.69 | N/A | N/A | N/A | N/A | N/A | N/A | 2.3 | 2.25 (0.59–8.57) | 0.23 |
| Other/noned | 1.1 | 1.64 (1.46–1.83) | <0.001 | 1.2 | 2.05 (1.61–2.60) | <0.001 | 0.5 | 3.49 (1.82–6.67) | <0.001 | 1.1 | 1.71 (0.97–3.01) | 0.06 | 0.9 | 1.62 (0.87–3.02) | 0.13 | 2.0 | 0.99 (0.35–2.75) | 0.98 | 3.8 | 1.16 (0.25–5.46) | 0.85 |
| BM after primary cancer diagnosis | NSCLC a (N = 2994) | SCLC (N = 839) | Breast (n = 1294) | Melanoma (N = 811) | Kidney (N = 351) | Colorectal (N = 560) | Esophagus (N = 157) | ||||||||||||||
| Median survival (mo.) | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | Median survival | HR (95% CI) | P value | |
| Initial BM management strategy, N (%) b | |||||||||||||||||||||
| Any non-stereotactic brain-directed radiation | 2.8 | Ref | -- | 4.0 | Ref | -- | 3.7 | Ref | -- | 2.3 | Ref | -- | 2.8 | Ref | -- | 2.1 | Ref | -- | 1.9 | Ref | -- |
| Any stereotactic brain-directed radiation | 7.9 | 0.56 (0.49–0.63) | <0.001 | 5.5 | 0.66 (0.48–0.92) | 0.01 | 8.6 | 0.60 (0.48–0.74) | <0.001 | 7.6 | 0.42 (0.33–0.53) | <0.001 | 7.3 | 0.58 (0.40–0.83) | 0.003 | 5.0 | 0.62 (0.47–0.82) | <0.001 | 3.3 | 0.74 (0.45–1.23) | 0.24 |
| Any neurosurgical resection | 8.3 | 0.50 (0.43–0.58) | <0.001 | 4.8 | 0.77 (0.37–1.63) | 0.49 | 7.7 | 0.66 (0.52–0.83) | <0.001 | 4.9 | 0.51 (0.40–0.66) | <0.001 | 18.6 | 0.35 (0.23–0.55) | <0.001 | 5.9 | 0.56 (0.42–0.75) | <0.001 | 7.0 | 0.40 (0.24–0.67) | <0.001 |
| Systemic therapy without local brain-directed therapyc | 5.4 | 0.68 (0.60–0.78) | <0.001 | 3.7 | 1.10 (0.90–1.35) | 0.33 | 6.2 | 0.65 (0.55–0.77) | <0.001 | 2.8 | 0.72 (0.53–0.98) | 0.03 | 2.9 | 0.84 (0.57–1.24) | 0.39 | 4.7 | 0.54 (0.37–0.77) | <0.001 | 1.3 | 0.77 (0.46–1.27) | 0.30 |
| Other/noned | 1.6 | 1.15 (1.05–1.26) | 0.003 | 2.1 | 1.13 (0.96–1.34) | 0.15 | 2.8 | 0.87 (0.75–1.01) | 0.06 | 1.4 | 0.97 (0.81–1.16) | 0.72 | 2.1 | 1.04 (0.79–1.36) | 0.79 | 1.7 | 0.87 (0.71–1.08) | 0.21 | 1.0 | 1.13 (0.72–1.78) | 0.59 |
Abbreviations: BM, brain metastasis; CI, confidence interval; HR, hazard ratio; N/A, not applicable because no patients in this treatment group; NSCLC, non-small-cell lung cancer; SCLC, small cell lung cancer.
aIncludes lung primaries that are not specifically listed as adenocarcinoma, squamous cell, adenosquamous, large cell, or bronchoalveolar (a histology still designated by SEER).
bPatients were considered recipients of a treatment if at least one of the relevant administrative codes (Table A1) was present within 1 month prior—2 months after BM diagnosis for brain-directed radiation or 1 month prior—1 month after BM diagnosis for neurosurgical resection.
cMay underestimate percent of patients receiving systemic therapy given that data on receipt of oral systemic agents were only available for patients with Part D coverage.
dMay include patients who received systemic therapy but for whom this information was not captured by claims due to lack of Part D coverage.
Among patients with metachronous BM, receipt of stereotactic brain-directed radiation was associated with improved survival compared with the reference of non-stereotactic brain-directed radiation among those with NSCLC (HR: 0.56 [0.49–0.63], P < 0.001), SCLC (HR: 0.66 [0.48–0.92], P = 0.01), melanoma (HR: 0.42 [0.33–0.53], P < 0.001), kidney (HR: 0.58 [0.40–0.83], P = 0.003), and colorectal (HR: 0.62 [0.47–0.82], P < 0.001) primary tumors. Neurosurgical resection was associated with improved survival among all primary sites (all P < 0.001) except for SCLC. Among patients with metachronous BM, systemic therapy without local brain-directed therapy was associated with improved survival for patients with NSCLC (HR: 0.68 [0.60–0.78], P < 0.001), breast (HR: 0.65 [0.55–0.77, P < 0.001), melanoma (HR: 0.72 [0.53–0.98], P = 0.03), and colorectal (HR: 0.54 [0.37–0.77], P < 0.001) primary tumors. Lack of treatment as captured by claims data was associated with poorer survival for patients with metachronous BM and NSCLC (HR: 1.15 [1.05–1.26], P = 0.003).
Discussion
In this large, population-based study, we characterized prognosis in older patients with solid malignancies who developed BM at or after primary cancer diagnosis. To our knowledge, ours is the first study to utilize SEER-Medicare data to characterize outcomes associated with BM that develop after primary cancer diagnosis, which comprised >70% of our cohort. While Cagney et al. recently published the first large-scale epidemiologic study of BM in the United States using the entire SEER database,13 reflecting a major contribution to the population-based literature on BM, this study was limited to patients with synchronous BM only given the confines of SEER data. We suspect that prior SEER-Medicare investigations relating to prognosis in patients with BM have been lacking given the fact that identification of a BM diagnosis date using claims data has only recently been validated.20 Here, we found survival for patients across all primary tumor sites to be poor, with median survival times of 4 months or less across all primaries except ovarian cancer. Notably, however, among patients with NSCLC receiving targeted therapy for presumed genetic alterations, survival was substantially longer. These results are highly generalizable given the data source utilized and are of potential utility to providers, health care systems, and policy efforts relating to BM, as well.
Prior efforts to characterize prognosis among patients with BM led to development of 2 commonly utilized prognostic indices, the RPA and subsequent DS-GPA.3,4,6,8,9 The RPA was based on analyses of 1200 patients with BM from 3 RTOG clinical trials between 1979 and 1993. The DS-GPA represented a more granular prognostic index, initially developed from patients enrolled in RTOG-led randomized clinical trials, and stratified patients into one of 4 prognostic categories; later efforts further stratified by disease site often using institutional/departmental data generally from larger academic centers. Median survivals for patients with BM in the original DS-GPA study were 7.00, 4.90, 11.93, and 6.74 months, for NSCLC, SCLC, breast, and melanoma, respectively, and represented a significant advance in prognostication among patients with BM.
Several studies have validated the DS-GPA as a prognostic tool for patients with BM.7–9 Subsequent studies have also attempted to further refine existing GPA indices via molecular and genetic characteristics,22 including EGFR and ALK alterations in NSCLC,25,26 receptor subtype in breast cancer,27 and BRAF status in melanoma.28 Such studies have demonstrated EGFR and ALK alterations, HER2-postivity, and BRAF mutations to be significant prognostic/predictive factors among NSCLC, breast, and melanoma patients with BM, respectively.27–30 Consistent with this literature, patients in our study receiving targeted therapy for presumed underlying mutations displayed longer survival than those without such alterations, although overall survival was still markedly lower than described in the DS- and molecular GPA studies (Supplementary Table 5),22,27,28,31–33 as well as compared with other dedicated efforts.34
The disparities in survival times between established prognostic indices and the SEER-Medicare data utilized in our study are striking and could be secondary to several important considerations. Firstly, patients in our study, with a median age of 75 at time of BM diagnosis, were significantly older than those in the GPA studies (median 60 y).6 In addition, patients in our study stemmed from a national registry, and consequently did not undergo the screening or selection process inherent to a clinical trial or institutionally/departmentally maintained database.6 Prior literature has demonstrated that clinical trial patients tend to be younger, healthier, and have a better baseline prognosis than non-trial patients.35–37 Consequently, extrapolation of trial results to real-world clinical practice becomes problematic. Moreover, later GPA analyses incorporated data from individual larger/academic centers, which may be prone to bias related to data collection; for example, radiation-based databases that primarily incorporate patients managed with radiosurgery, a cohort of significant academic interest, can impart both immortal time bias38 and confounding by indication. Indeed, in our study, patients managed with radiosurgery displayed better survival than patients managed via other approaches.
The guarded prognosis for elderly patients with BM has important implications for management. If prognosis is driven by systemic and not intracranial disease burden, avoidance of brain-directed, potentially toxic treatments, whole-brain radiotherapy (WBRT) in particular, seems prudent. Such an approach is consistent with the conclusion of the QUARTZ trial,39 which randomized patients with NSCLC and BM unsuitable for neurosurgical resection or stereotactic radiation therapy to best supportive care or WBRT; ultimately, no difference in overall survival was found between the 2 groups. The median survival in both cohorts was approximately 2 months, likely driven by extracranial as opposed to intracranial disease burden given that most patients had uncontrolled primary lesions and harbored extracranial metastases, while only approximately one-third had >5 BM at enrollment. The authors ultimately concluded that WBRT provides little benefit, if any, for this patient group. In our cohort, a substantial proportion of patients also had extracranial metastases and had progressed through multiple systemic agents, suggestive of a significant extracranial disease burden at BM diagnosis. Multiple prior studies have demonstrated that the competing risk of non-neurologic death due to systemic disease progression is greater than the risk of neurologic death among such patients.40–42 Consistent with this work, we found that regardless of primary tumor site, patients managed with non-stereotactic brain-directed radiation therapy (encompassing WBRT) had median survival times between 2 to 4 months only. Although patients managed with stereotactic brain-directed radiation therapy or neurosurgical resection appeared to have better outcomes among most primary tumor sites compared with patients who received non-stereotactic brain-directed radiation therapy (Table 6), their improved survival likely reflects selection bias as opposed to treatment efficacy given that such treatments are generally utilized for patients with a lower burden of intracranial disease, and that omission of WBRT in lieu of therapy such as stereotactic radiation or resection has generally not shown improvements in overall survival based on prior randomized studies.43–50 However, brain-directed stereotactic radiation is typically associated with a very favorable short-term adverse effect profile, provides excellent local control of treated metastases, and often can palliate or prevent neurologic symptomatology due to intracranial disease. Therefore, stereotactic brain-directed radiation may reflect an optimal management strategy for older patients with a limited or moderate intracranial disease burden given the generally guarded overall prognosis that these patients display.
Our work should be considered in the context of its limitations. Firstly, claims data cannot be reliably utilized to identify metastatic involvement of many sites, and the National Cancer Institute advises caution when using claims to identify metastases after primary cancer diagnosis.51 However, because BM, unlike most distant sites, are commonly treated with local therapies for which diagnostic/billing codes exist, claims data can be used to reliably identify BM. In fact, the use of health insurance claims data for identification of BM has been validated with manual chart review and shown to have a high sensitivity (>97%) and specificity (99%).19,20 Secondly, we could not directly differentiate between non-neurologic and neurologic causes of death, although we suspect based on prior literature that most patients died of systemic disease progression.40–42 Thirdly, we did not have many of the granular variables available in the GPA studies, including performance status, number of BM, and extracranial disease burden, which all have important implications for predicting individual patient prognosis. Lastly, 2016 is the most recent year for which SEER-Medicare data are available, and therefore, the impact of newer systemic agents could not be assessed.
Conclusions
In this population-based study of nearly 10 000 elderly patients with synchronous and metachronous BM, we identified a substantially poorer prognosis than previously described. The results of our study suggest that many elderly patients may derive less benefit from aggressive intracranial treatments with potential for short and long-term toxicity, especially if extracranial disease burden is significant and/or limited systemic options exist. For such patients, management strategies that promote quality of life and symptom palliation with minimal morbidity should be prioritized.
Supplementary Material
Acknowledgments
This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the National Cancer Institute; the Office of Research, Development and Information, CMS; Information Management Services (IMS), Inc; and the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the SEER-Medicare database.
Funding
No funding was required for this study.
Conflict of interest statement. Dr Aizer reports research funding from Varian Medical Systems and consulting fees from Novartis. The remaining authors declare no conflicts of interest.
Authorship statement. Study conception/design: NL, AAA. Data collection/analysis/interpretation: All authors. Statistical analysis: NL, PJC, AAA. Drafting of the manuscript: NL, AAA. Manuscript editing/critical revision of the manuscript: all authors. Supervision: AAA.
References
- 1. Nayak L, Lee EQ, Wen PY. Epidemiology of brain metastases. Curr Oncol Rep. 2012;14(1):48–54. [DOI] [PubMed] [Google Scholar]
- 2. Achrol AS, Rennert RC, Anders C, et al. . Brain metastases. Nat Rev Dis Primers. 2019;5(1):5. [DOI] [PubMed] [Google Scholar]
- 3. Gaspar L, Scott C, Rotman M, et al. . Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys. 1997;37(4):745–751. [DOI] [PubMed] [Google Scholar]
- 4. Sperduto PW, Chao ST, Sneed PK, et al. . Diagnosis-specific prognostic factors, indexes, and treatment outcomes for patients with newly diagnosed brain metastases: a multi-institutional analysis of 4259 patients. Int J Radiat Oncol Biol Phys. 2010;77(3):655–661. [DOI] [PubMed] [Google Scholar]
- 5. Kavouridis VK, Harary M, Hulsbergen AFC, et al. . Survival and prognostic factors in surgically treated brain metastases. J Neurooncol. 2019;143(2):359–367. [DOI] [PubMed] [Google Scholar]
- 6. Sperduto PW, Berkey B, Gaspar LE, Mehta M, Curran W. A new prognostic index and comparison to three other indices for patients with brain metastases: an analysis of 1960 patients in the RTOG database. Int J Radiat Oncol Biol Phys. 2008;70(2):510–514. [DOI] [PubMed] [Google Scholar]
- 7. Berghoff AS, Schur S, Füreder LM, et al. . Descriptive statistical analysis of a real life cohort of 2419 patients with brain metastases of solid cancers. ESMO Open. 2016;1(2):e000024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Gaspar LE, Scott C, Murray K, Curran W. Validation of the RTOG recursive partitioning analysis (RPA) classification for brain metastases. Int J Radiat Oncol Biol Phys. 2000;47(4):1001–1006. [DOI] [PubMed] [Google Scholar]
- 9. Sperduto CM, Watanabe Y, Mullan J, et al. . A validation study of a new prognostic index for patients with brain metastases: the Graded Prognostic Assessment. J Neurosurg. 2008;109(Suppl):87–89. [DOI] [PubMed] [Google Scholar]
- 10. Agboola O, Benoit B, Cross P, et al. . Prognostic factors derived from recursive partition analysis (RPA) of Radiation Therapy Oncology Group (RTOG) brain metastases trials applied to surgically resected and irradiated brain metastatic cases. Int J Radiat Oncol Biol Phys. 1998;42(1):155–159. [DOI] [PubMed] [Google Scholar]
- 11. Cannady SB, Cavanaugh KA, Lee SY, et al. . Results of whole brain radiotherapy and recursive partitioning analysis in patients with brain metastases from renal cell carcinoma: a retrospective study. Int J Radiat Oncol Biol Phys. 2004;58(1):253–258. [DOI] [PubMed] [Google Scholar]
- 12. Videtic GM, Adelstein DJ, Mekhail TM, et al. . Validation of the RTOG recursive partitioning analysis (RPA) classification for small-cell lung cancer-only brain metastases. Int J Radiat Oncol Biol Phys. 2007;67(1):240–243. [DOI] [PubMed] [Google Scholar]
- 13. Cagney DN, Martin AM, Catalano PJ, et al. . Incidence and prognosis of patients with brain metastases at diagnosis of systemic malignancy: a population-based study. Neuro Oncol. 2017;19(11): 1511–1521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pestalozzi BC, Holmes E, de Azambuja E, et al. . CNS relapses in patients with HER2-positive early breast cancer who have and have not received adjuvant trastuzumab: a retrospective substudy of the HERA trial (BIG 1-01). Lancet Oncol. 2013;14(3):244–248. [DOI] [PubMed] [Google Scholar]
- 15. Kelly E, Stephens P. Cancer in the older person. Rev Clin Gerontol. 2015;25(3):172–180. [Google Scholar]
- 16. Minniti G, Filippi AR, Osti MF, Ricardi U. Radiation therapy for older patients with brain tumors. Radiat Oncol. 2017. doi: 10.1186/s13014-017-0841-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Surveillance, Epidemiology, and End Results (SEER) Program; National Cancer Institute. Overview of the SEER Program.https://seer.cancer.gov/about/overview.html. Accessed September 16, 2019.
- 18. National Cancer Institute, Division of Cancer Control and Population Sciences. SEER-Medicare: How the SEER & Medicare Data are Linked. https://healthcaredelivery.cancer.gov/seermedicare/overview/linked.html. Accessed September 16, 2019.
- 19. Eichler AF, Lamont EB. Utility of administrative claims data for the study of brain metastases: a validation study. J Neurooncol. 2009;95(3):427–431. [DOI] [PubMed] [Google Scholar]
- 20. Lamba N, Kearney RB, Mehanna E, et al. . Utility of claims data for identification of date of diagnosis of brain metastases. Neuro Oncol. 2020. doi: 10.1093/neuonc/noz245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Lam MB, Li L, Cronin A, Schrag D, Chen AB. Palliative radiation and fractionation in medicare patients with incurable non-small cell lung cancer. Adv Radiat Oncol. 2018;3(3):382–390. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sperduto PW, Yang TJ, Beal K, et al. . Estimating survival in patients with lung cancer and brain metastases: an update of the graded prognostic assessment for Lung Cancer Using Molecular Markers (Lung-molGPA). JAMA Oncol. 2017;3(6):827–831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Dummer R, Schadendorf D, Ascierto PA, et al. . Binimetinib versus dacarbazine in patients with advanced NRAS-mutant melanoma (NEMO): a multicentre, open-label, randomised, phase 3 trial. Lancet Oncol. 2017;18(4):435–445. [DOI] [PubMed] [Google Scholar]
- 24. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–619. [DOI] [PubMed] [Google Scholar]
- 25. Sperduto PW, Yang TJ, Beal K, et al. . The effect of gene alterations and tyrosine kinase inhibition on survival and cause of death in patients with adenocarcinoma of the lung and brain metastases. Int J Radiat Oncol Biol Phys. 2016;96:406–413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Park SJ, Kim HT, Lee DH, et al. . Efficacy of epidermal growth factor receptor tyrosine kinase inhibitors for brain metastasis in non-small cell lung cancer patients harboring either exon 19 or 21 mutation. Lung Cancer. 2012;77(3):556–560. [DOI] [PubMed] [Google Scholar]
- 27. Sperduto PW, Kased N, Roberge D, et al. . Effect of tumor subtype on survival and the graded prognostic assessment for patients with breast cancer and brain metastases. Int J Radiat Oncol Biol Phys. 2012;82(5):2111–2117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Sperduto PW, Jiang W, Brown PD, et al. . Estimating survival in melanoma patients with brain metastases: an update of the graded prognostic assessment for Melanoma Using Molecular Markers (Melanoma-molGPA). Int J Radiat Oncol Biol Phys. 2017;99(4):812–816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Nieder C, Hintz M, Oehlke O, Bilger A, Grosu AL. Validation of the graded prognostic assessment for lung cancer with brain metastases using molecular markers (lung-molGPA). Radiat Oncol. 2017;12(1):107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Chen K, Yu X, Zhang F, et al. . Applicability of the lung-molGPA index in non-small cell lung cancer patients with different gene alterations and brain metastases. Lung Cancer. 2018;125:8–13. [DOI] [PubMed] [Google Scholar]
- 31. Sperduto PW, Kased N, Roberge D, et al. . 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. 2012. doi: 10.1200/JCO.2011.38.0527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Sperduto PW, Fang P, Li J, et al. . Estimating survival in patients with gastrointestinal cancers and brain metastases: An update of the graded prognostic assessment for gastrointestinal cancers (GI-GPA). Clin Transl Radiat Oncol. 2019;18:39–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Sperduto PW, Deegan BJ, Li J, et al. . Estimating survival for renal cell carcinoma patients with brain metastases: an update of the Renal Graded Prognostic Assessment tool. Neuro Oncol. 2018;20(12):1652–1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Johung KL, Yeh N, Desai NB, et al. . Extended survival and prognostic factors for patients with ALK-rearranged non-small-cell lung cancer and brain metastasis. J Clin Oncol. 2016;34(2):123–129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Khozin S, Blumenthal GM, Pazdur R. Real-world data for clinical evidence generation in oncology. J Natl Cancer Inst. 2017;109(11). doi: 10.1093/jnci/djx187. [DOI] [PubMed] [Google Scholar]
- 36. Kennedy-Martin T, Curtis S, Faries D, Robinson S, Johnston J. A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials. 2015;16:495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Karanis YB, Canta FAB, Mitrofan L, Mistry H, Anger C. ‘Research’ vs ‘real world’ patients: the representativeness of clinical trial participants. Ann Oncol. 2016;27:vi542. [Google Scholar]
- 38. Park HS, Gross CP, Makarov DV, Yu JB. Immortal time bias: a frequently unrecognized threat to validity in the evaluation of postoperative radiotherapy. Int J Radiat Oncol Biol Phys. 2012;83(5):1365–1373. [DOI] [PubMed] [Google Scholar]
- 39. Mulvenna P, Nankivell M, Barton R, et al. . Dexamethasone and supportive care with or without whole brain radiotherapy in treating patients with non-small cell lung cancer with brain metastases unsuitable for resection or stereotactic radiotherapy (QUARTZ): results from a phase 3, non-inferiority, randomised trial. Lancet. 2016;388(10055):2004–2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Yamamoto M, Serizawa T, Shuto T, et al. . Stereotactic radiosurgery for patients with multiple brain metastases (JLGK0901): A multi-institutional prospective observational study. Lancet Oncol. 2014;15(4):387–395. [DOI] [PubMed] [Google Scholar]
- 41. McTyre ER, Johnson AG, Ruiz J, et al. . Predictors of neurologic and nonneurologic death in patients with brain metastasis initially treated with upfront stereotactic radiosurgery without whole-brain radiation therapy. Neuro Oncol. 2017;19(4):558–566. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Lucas JT Jr, Colmer HG 4th, White L, et al. . Competing risk analysis of neurologic versus nonneurologic death in patients undergoing radiosurgical salvage after whole-brain radiation therapy failure: who actually dies of their brain metastases? Int J Radiat Oncol Biol Phys. 2015;92(5):1008–1015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Brown PD, Jaeckle K, Ballman KV, et al. . Effect of radiosurgery alone vs radiosurgery with whole brain radiation therapy on cognitive function in patients with 1 to 3 brain metastases: a randomized clinical trial. JAMA. 2016;316(4):401–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Chang EL, Wefel JS, Hess KR, et al. . Neurocognition in patients with brain metastases treated with radiosurgery or radiosurgery plus whole-brain irradiation: a randomised controlled trial. Lancet Oncol. 2009;10(11):1037–1044. [DOI] [PubMed] [Google Scholar]
- 45. Aoyama H, Shirato H, Tago M, et al. . Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial. JAMA. 2006;295(21):2483–2491. [DOI] [PubMed] [Google Scholar]
- 46. Brown PD, Ballman KV, Cerhan JH, et al. . Postoperative stereotactic radiosurgery compared with whole brain radiotherapy for resected metastatic brain disease (NCCTG N107C/CEC·3): a multicentre, randomised, controlled, phase 3 trial. Lancet Oncol. 2017;18(8):1049–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Kocher M, Soffietti R, Abacioglu U, et al. . Adjuvant whole-brain radiotherapy versus observation after radiosurgery or surgical resection of one to three cerebral metastases: results of the EORTC 22952-26001 study. J Clin Oncol. 2011;29(2):134–141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Patchell RA, Tibbs PA, Regine WF, et al. . Postoperative radiotherapy in the treatment of single metastases to the brain: a randomized trial. JAMA. 1998;280(17):1485–1489. [DOI] [PubMed] [Google Scholar]
- 49. Kayama T, Sato S, Sakurada K, et al. . Effects of surgery with salvage stereotactic radiosurgery versus surgery with whole-brain radiation therapy in patients with one to four brain metastases (JCOG0504): A Phase III, noninferiority, randomized controlled trial. J Clin Oncol. 2018;36(33):3282–3289. [DOI] [PubMed] [Google Scholar]
- 50. Kępka L, Tyc-Szczepaniak D, Bujko K, et al. . Stereotactic radiotherapy of the tumor bed compared to whole brain radiotherapy after surgery of single brain metastasis: Results from a randomized trial. Radiother Oncol. 2016;121(2):217–224. [DOI] [PubMed] [Google Scholar]
- 51. National Cancer Institute. SEER-Medicare Linked Database. https://healthcaredelivery.cancer.gov/seermedicare/considerations/measures.html#13. Accessed November 1, 2019.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.