Abstract
Background
Ependymoma is a rare CNS tumor arising from the ependymal lining of the ventricular system. General differences in incidence and survival have been noted but not examined on a comprehensive scale for all ages and by histology. Despite the rarity of ependymomas, morbidity/mortality associated with an ependymoma diagnosis justifies closer examination.
Methods
Incidence data were obtained from the Central Brain Tumor Registry of the United States in collaboration with the Centers for Disease Control and Prevention and the National Cancer Institute, and survival data from Surveillance Epidemiology and End Results, from 2000 to 2016 for anaplastic ependymoma and ependymoma, not otherwise specified (NOS). Age-adjusted incidence rates (IRs) per 100 000 person-years were analyzed by age, sex, race, and location. Survival analysis was performed with Kaplan-Meier curves and multivariable Cox proportional hazards models.
Results
Incidence of anaplastic ependymoma was highest in ages 0 to 4 years. African American populations had lower incidence but had a 78% increased risk of death compared to white populations (hazard ratio [HR]: 1.78 [95% CI, 1.30-2.44]). Incidence was highest for anaplastic ependymoma in the supratentorial region. Adults (age 40+ years) had almost twice the risk of death compared to children (ages 0-14 years) (HR: 1.97 [95% CI, 1.45-2.66]). For ependymoma, NOS, subtotal resection had a risk of mortality 1.86 times greater than gross total resection ([HR: 1.86 [95% CI, 1.32-2.63]).
Conclusions
African American populations experienced higher mortality rates despite lower incidence compared to white populations. Extent of resection is an important prognostic factor for survival. This highlights need for further evaluation of treatment patterns and racial disparities in the care of patients with ependymoma subtypes.
Keywords: CBTRUS, ependymoma, incidence, survival
Ependymoma is a rare CNS tumor that is thought to arise from the radial glial cells, precursors of ependymal cells lining the ventricular system.1 Treatment of ependymoma varies by tumor location, but typically involves maximum, safe surgical resection with or without postoperative radiation therapy. There is no standard chemotherapy treatment for ependymoma, although some patients may benefit from the combination of lapatinib and temozolomide at recurrence.2 Prior epidemiological studies have noted trends in ependymoma incidences overall. For example, from 2012 to 2016, 0.29 individuals per 100 000 between ages 0 and 19 years were diagnosed with ependymoma.3 Additionally, the overall incidence of ependymoma increased from 2000 to 2010 with an annual percent change (APC) of 2.1%.4 Other previous studies highlighted the increased incidence ratio for male compared to female patients, but found conflicting results in terms of incidence of anaplastic ependymoma in children and adolescents compared to adults.4–7 Few studies, however, have examined the incidence of ependymoma on a comprehensive scale for all ages and specific histologic subtypes, such as anaplastic ependymoma. Therefore, because of the rarity of ependymomas in comparison to other CNS tumors, the increasing incidence rates (IRs) overall and the morbidity/mortality associated with ependymoma diagnosis justify a closer examination of these subtypes to better characterize nationwide trends.
In addition to changing IRs, few studies have analyzed recent population-based survival trends for patients diagnosed with different types of ependymoma. Amirian et al found age, tumor site, extent of resection, and tumor histology to be significant predictors of prognosis in individuals diagnosed with ependymoma, with a supratentorial tumor location being the strongest predictor of poor outcome.8 Additional studies that examined ependymoma survival based on factors such as histology and tumor location or specific age groups found varying survival trends.9–12 One study identified having gross-total resection (GTR) as the only predictor of progression-free survival, whereas other studies have shown patient sex and CNS location (posterior fossa vs supratentorial) to be the most significant predictors of survival.13–15 The aim of this study, therefore, is to provide an updated and comprehensive overview of incidence rates and survival trends for 2 types of ependymoma, anaplastic ependymoma, and ependymoma not otherwise specified (NOS), in the United States because these are the most commonly reported ependymoma histologies in the Central Brain Tumor Registry of the United States (CBTRUS).
Methods
This study was approved by the University Hospitals Cleveland Medical Center Institutional Review Board. Data obtained from the CBTRUS, in collaboration with the Centers for Disease Control and Prevention (CDC) and National Cancer Institute (NCI), were obtained from central cancer registries from 50 state cancer registries and the registry of the District of Columbia. The final data set included 47 CDC National Program of Cancer Registration (NPCR) and 4 NCI Surveillance Epidemiology and End Results (SEER) central cancer registries, and provides incidence data representing the US population. Anaplastic ependymomas were identified by the International Classification of Diseases for Oncology, third edition (ICD-O-3) code: 9392/3. Ependymoma, NOS was identified by ICD-O-3 code 9391/3. Information on patient survival outcomes was obtained from SEER data and was analyzed to generate survival data both for anaplastic ependymoma and ependymoma, NOS from 2000 to 2016. In this data set, the SEER registry system consists of 18 registries representing a subset of the population included in the CBTRUS dataset. According to the US 2000 Census, data pulled from the 18 SEER registries provide population-based information for approximately 28% of the US population.
Age-adjusted IRs were obtained for anaplastic ependymoma and ependymoma, NOS from 2000 to 2016 and were standardized to the 2000 US standard population and reported per 100 000 population. IRs were calculated using SEER*Stat 8.3.6, and figures were generated using R version 3.5.2 (R Core Team, 2018). Descriptive statistics were collected, and chi-square tests performed to assess differences in proportions (Table 1). Statistics were excluded for cells containing fewer than 16 counts as required by NPCR given the poor reliability of rates computed with small numbers (https://www.cdc.gov/cancer/uscs/technical_notes/stat_methods/suppression.htm). Age-adjusted IRs and 95% CI were estimated for anaplastic ependymoma and ependymoma, NOS by sex, race, ethnicity, and tumor location including supratentorial (ICD-O-3 codes C700 and C709-C714), infratentorial (C716-C717), and spinal (C701, C720-C721, and C725). IRs were also presented by age. Patients with primary brain and CNS ependymoma in other nonspecified sites were excluded from location-specific IR comparisons but were included in nonlocational-specific incidence. Race categories for this study included white, African American, American Indian/Alaskan Native, and Asian/Pacific Islander (API). Incidence for Hispanic patients vs non-Hispanic patients was also analyzed. Unknown, unspecified, and other race categories were excluded from race-specific IR comparisons but were included in non–race-specific incidence. Joinpoint Regression Program 4.6.0.0 software was used to compute APC in IRs from 2000 to 2016 to examine trends over time. Joinpoint software selects a minimum number of joinpoints to prohibit statistically significant improvement if one additional joinpoint is added (http://surveillance.cancer.gov/joinpoint).
Table 1.
Descriptive Statistics for Patients Diagnosed With Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified (NOS), Central Brain Tumor Registry of the United States 2000-2016a
| Histology | ||||
|---|---|---|---|---|
| Overall | Ependymoma, NOS | Anaplastic ependymoma | P | |
| No. | 13 123 (100%) | 10 991 (83.8%) | 2132 (16.2%) | |
| Age group, no. (%), y | < .001 | |||
| 0-4 | 1468 (11.2%) | 757 (6.9%) | 711 (33.3%) | |
| 5-9 | 700 (5.3%) | 422 (3.8%) | 278 (13.0%) | |
| 10-14 | 572 (4.4%) | 398 (3.6%) | 174 (8.2%) | |
| 15-19 | 552 (4.2%) | 416 (3.8%) | 136 (6.4%) | |
| 20-24 | 548 (4.2%) | 448 (4.1%) | 100 (4.7%) | |
| 25-29 | 708 (5.4%) | 628 (5.7%) | 80 (3.8%) | |
| 30-34 | 760 (5.8%) | 704 (6.4%) | 56 (2.6%) | |
| 35-39 | 903 (6.9%) | 823 (7.5%) | 80 (3.8%) | |
| 40-44 | 1037 (7.9%) | 949 (8.6%) | 88 (4.1%) | |
| 45-49 | 1146 (8.7%) | 1058 (9.6%) | 88 (4.1%) | |
| 50-54 | 1172 (8.9%) | 1085 (9.9%) | 87 (4.1%) | |
| 55-59 | 1052 (8.0%) | 980 (8.9%) | 72 (3.4%) | |
| 60-64 | 784 (6.0%) | 732 (6.7%) | 52 (2.4%) | |
| 65-69 | 662 (5.0%) | 614 (5.6%) | 48 (2.3%) | |
| 70-74 | 491 (3.7%) | 452 (4.1%) | 39 (1.8%) | |
| 75-79 | 336 (2.6%) | 314 (2.9%) | 22 (1.0%) | |
| 80-84 | 164 (1.2%) | – | – | |
| 85+ | 68 (0.5%) | – | – | |
| Race | < .001 | |||
| American Indian/Alaska Native | 92 (0.7%) | 75 (0.7%) | 17 (0.8%) | |
| Asian or Pacific Islander | 434 (3.4%) | 338 (3.1%) | 96 (4.6%) | |
| African American | 1189 (9.2%) | 942 (8.7%) | 247 (11.7%) | |
| White | 11 230 (86.8%) | 9485 (87.5%) | 1745 (82.9%) | |
| Ethnicity, no. (%) | < .001 | |||
| Non–Spanish-Hispanic-Latino | 11 245 (85.7%) | 9510 (86.5%) | 1735 (81.4%) | |
| Spanish-Hispanic-Latino | 1878 (14.3%) | 1481 (13.5%) | 397 (18.6%) | |
| Location, No. (%) | < .001 | |||
| Infratentorial | 2619 (20.0%) | 2100 (19.1%) | 519 (24.3%) | |
| Spinal | 6049 (46.1%) | 5878 (53.5%) | 171 (8.0%) | |
| Supratentorial | 1505 (11.5%) | 695 (6.3%) | 810 (38.0%) | |
| Other/Unspecified | 2950 (22.5%) | 2318 (21.1%) | 632 (29.6%) | |
| Surgery | < .001 | |||
| Gross total resection | 5010 (38.2%) | 1080 (50.7%) | 3930 (35.8%) | |
| Subtotal resection | 1861 (14.2%) | 404 (18.9%) | 1457 (13.3%) | |
| Biopsy | 2872 (21.9%) | 425 (19.9%) | 2447 (22.3%) | |
| None | 1113 (8.5%) | 82 (3.8%) | 1031 (9.4%) | |
| Other/Unknown | 2267 (17.3%) | 141 (6.6%) | 2126 (19.3%) | |
| Radiation | < .001 | |||
| Yes | 4266 (32.5%) | 1429 (67.0%) | 2837 (25.8%) | |
| No/Unknown | 8857 (67.5%) | 703 (33.0%) | 8154 (74.2%) | |
aData provided to the Central Brain Tumor Registry of the United States in collaboration with the Centers for Disease Control and Prevention and National Cancer Institute.
–Sites with fewer than 16 cases are suppressed.
Survival differences were assessed among age, race/ethnicity, sex, location, and treatment modality (GTR, subtotal resection [STR], GTR + radiotherapy (RT), and STR + RT). These survival differences were assessed with Kaplan-Meier survival curves and multivariable Cox proportional hazards models. Kaplan-Meier survival curves were plotted, and the survival differences were assessed using the log-rank test and multivariate Cox proportional hazards models. R version 3.5.2 (R Core Team, 2018) was used to perform all statistical analyses. The Cox proportional hazards models included all a priori specified covariates; age, sex, race, ethnicity, histology, location, and treatment, regardless of individual significance level. Level for statistical significance was set at P less than .05.
Results
Descriptive Statistics
Demographic and treatment comparison between anaplastic ependymoma and ependymoma, NOS are depicted in Table 1. There was a notable difference in age distribution, with the highest proportion of anaplastic ependymoma patients being age 0 to 4 years (33.3%), as opposed to ependymoma, NOS, which had highest proportion among those age 50 to 54 years (9.9%). There was also a large difference in location, with most anaplastic ependymoma cases being located in the supratentorial region (38.0%), whereas most ependymoma, NOS cases occurred in the spinal cord (53.5%). Ependymoma, NOS patients also had a higher proportion of GTR, (50.7%) and radiation therapy (RT) (67.0%) compared to anaplastic ependymoma (GTR: 35.8%, RT: 25.8%).
Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified Incidence by Age From 2000 to 2016
IR and incidence trends for anaplastic ependymoma and ependymoma, NOS from 2000 to 2016 are depicted in Figures 1 and 2, and Supplemental Figure 1. The overall incidence of anaplastic ependymoma was higher in age 0 to 4 years (0.43 cases per 100 000 [95% CI, 0.41-0.46]) compared to all other age groups. The overall incidence in age 5 to 9 years was higher when compared to older age groups as well. After age 20 years, incidence of anaplastic ependymoma remained generally stable (Figure 1A). The overall incidence of ependymoma NOS, however, was elevated in age 0 to 4 years at 0.22 cases per 100 000 (95% CI, 0.21-0.24) but also peaked at 0.31 cases per 100 000 (95% CI, 0.29-0.33) in individuals age 55 to 59 years, demonstrating a bimodal distribution (Figure 2A). From 2000 to 2016 there was a significant increase in anaplastic ependymoma incidence for individuals age 0 to 4 years (APC: 5.91% [95% CI, 1.55-10.45]) and for individuals age 5 to 9 years (APC: 3.83% [95% CI, 1.46-6.27]) (Figure 1B). There was a significant decrease, however, in incidence of ependymoma, NOS in individuals age 0 to 4 years (APC: –4.00% [95% CI, –4.93 to –3.06]) and in individuals age 5 to 9 years (APC: –3.15% [95% CI, –5.23 to –1.02]) (Figure 2B). Interestingly, there was no significant difference in the incidence of anaplastic ependymoma and ependymoma, NOS in ages 0 to 4 years. This is in contrast to all other pediatric age groups, in which ependymoma, NOS incidence was significantly higher (Figures 1A and 2A).
Figure 1.
A, Age-adjusted incidence rates for anaplastic ependymoma by 5-year age groupings. B, Age-adjusted incidence rates for anaplastic ependymoma over time, by age group. C, Age-adjusted incidence rates for anaplastic ependymoma over time, by sex. D, Age-adjusted incidence rates for anaplastic ependymoma over time, by race. E, Age-adjusted incidence rates for anaplastic ependymoma over time, by ethnicity. F, Age-adjusted incidence rates for anaplastic ependymoma over time, by location. *Significant annual percent change (Central Brain Tumor Registry of the United States 2000-2016).
Figure 2.
A, Age-adjusted incidence rates for ependymoma, not otherwise specified (NOS) by 5-year age groupings. B, Age-adjusted incidence rates for ependymoma, NOS over time, by age group. C, Age-adjusted incidence rates for ependymoma, NOS over time, by sex. D, Age-adjusted incidence rates for ependymoma, NOS over time, by race. E, Age-adjusted incidence rates for ependymoma, NOS over time, by ethnicity. F, Age-adjusted incidence rates for ependymoma, NOS over time, by location. *Significant annual percent change (Central Brain Tumor Registry of the United States 2000-2016).
Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified Incidence by Sex From 2000 to 2016
Age-adjusted IR and incidence trends for anaplastic ependymoma and ependymoma, NOS for male and female patients from 2000 to 2016 are depicted in Figure 1C and 2C, and Supplemental Figure 1A. The overall incidence of anaplastic ependymoma increased both in male (APC: 5.56% [95% CI, 3.89-7.25]) and female patients (APC: 3.45% [95% CI, 1.95-4.97]) (Figure 1A). The incidence of ependymoma, NOS in male patients increased significantly from 2000 to 2005 (APC: 4.3% [95% CI, 0.84-7.89]), but subsequently decreased from 2005 to 2016 (APC: –1.25% [95% CI, –1.84 to –0.66]). Likewise, the incidence of ependymoma, NOS in female patients increased significantly from 2000 to 2004 (APC: 7.86% [95% CI, 1.65-14.45]), but decreased from 2004 to 2016 (APC: –2.79% [95% CI, –3.83 to –1.74]) (Figure 2C). For male patients, the incidence of ependymoma, NOS (0.225 per 100 000 [95% CI, 0.219-0.231]) was greater than incidence in female patients (0.20 per 100 000 [95% CI, 0.194-0.205]) (Supplemental Figure 1A). However, there was no significant difference in anaplastic ependymoma incidence between male and female patients (0.045 per 100 000 [95% CI, 0.042-0.047] vs 0.040 per 100 000 [95% CI, 0.040-0.042], respectively) (Supplemental Figure 1A).
Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified Incidence by Race and Ethnicity from 2000 to 2016
Age-adjusted IR and incidence trends for anaplastic ependymoma and ependymoma, NOS from 2000 to 2016 are depicted by race in Figures 1D and 2D, and Supplemental Figure 1B, and ethnicity in Figures 1E and 2E, and Supplemental Figure 1C. The IR of anaplastic ependymoma increased significantly from 2000 to 2016 both for white (APC: 4.02% [95% CI, 2.62-5.44]) and African American (APC: 4.16% [95% CI, 0.41-8.04]) populations (Figure 1D). The IR of anaplastic ependymoma increased significantly both for Hispanic (APC: 3.53% [95% CI, 0.96-6.16]) and non-Hispanic populations (APC: 4.17% (95% CI, 2.95-5.41]) from 2000 to 2016 as well (Figure 1E). The IR of ependymoma, NOS among whites increased significantly from 2000 to 2004 (APC: 5.57% [95% CI, 2.45-8.78]), but decreased significantly from 2004 to 2016 (APC: –1.86% [95% CI, –2.39 to –1.32]). The IR of ependymoma, NOS among African Americans did not vary significantly from 2000 to 2016 (Figure 2D). IRs of ependymoma, NOS increased significantly for non-Hispanic populations from 2000 to 2005 (APC: 4.16% (95% CI, 0.93-7.49]) but decreased significantly from 2005 to 2016 (APC: –2.15% (95% CI, –3.08 to –1.21]). The IR of ependymoma, NOS among Hispanic populations, however, did not vary significantly (Figure 2E). Among white populations, IRs both of anaplastic ependymoma (0.04 per 100 000 [95% CI, 0.04-0.46]) and ependymoma, NOS (0.23 per 100 000 [95% CI, 0.222-0.23]) were significantly greater than all other races except for anaplastic ependymoma in API populations (Supplemental Figure 1B). IRs for anaplastic ependymoma did not vary between Hispanic and non-Hispanic populations. IRs for ependymoma, NOS, however, was slightly greater among non-Hispanic individuals (0.215 per 100 000 [95% CI, 0.210-0.22]) when compared to Hispanic individuals (0.198 per 100 000 [95% CI, 0.187-0.209]) (Supplemental Figure 1C).
Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified Incidence by Location from 2000 to 2016
Age-adjusted IRs and incidence trends for anaplastic ependymoma and ependymoma, NOS from 2000 to 2016 are depicted by location in Figure 1F and 2F, and Supplemental Figure 1D. The majority of anaplastic ependymomas were located in the supratentorial brain regions (0.016 per 100 000 [95% CI, 0.015-0.017]), followed closely by the infratentorial brain regions (0.010 per 100 000 [95% CI, 0.009-0.011]) and spine (0.003 per 100 000 [95% CI, 0.0026-0.0037]) (Supplemental Figure 1D). From 2000 to 2016, the rate of supratentorial (APC: 3.00% [95% CI, 1.18-4.85]), infratentorial (APC: 5.29% [95% CI, 2.84-7.79]), and spinal (APC: 5.01% [95% CI, 1.59-8.54]) anaplastic ependymoma increased significantly (Figure 1F).
In contrast, the majority of ependymoma, NOS were located in the spine (0.111 per 100 000 [95% CI, 0.108-0.114]), followed by the infratentorial brain (0.04 per 100 000 [95% CI, 0.038-0.042]) and supratentorial brain, respectively (0.014 per 100 000 [95% CI, 0.013-0.015]) (Supplemental Figure 1D). The rate of supratentorial ependymoma NOS increased (APC: 17.97% [95% CI, 0.08-39.06]) from 2000 to 2003, but subsequently decreased from 2003 to 2016 (APC: –4.64% [95% CI, –6.27 to –2.99]). The rate of spinal ependymoma, NOS decreased significantly from 2004 to 2016 (APC: –1.57% [95% CI, –2.55 to –0.58]) (Figure 2F).
Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified Survival by Age, Race/Ethnicity, Histology, Radiation, Sex, Location, and Treatment Modality from 2000 to 2016
Kaplan-Meier estimates for grouped anaplastic ependymoma and ependymoma, NOS showed significant differences in overall survival by age (P < .001), race (P < .001), histology (P < .001), sex (P = .002), and location (P < .001) (Supplemental Figure 2). Overall survival for anaplastic ependymoma and for ependymoma, NOS were significantly affected by resection type (STR vs GTR) and combined RT (anaplastic: P < .001, NOS: P < .001) with GTR conferring a survival advantage over STR regardless of adjuvant RT (Figures 3A-3C).
Figure 3.
A, Kaplan-Meier survival curves depicting the differences in survival for anaplastic ependymoma and ependymoma, not otherwise specified (NOS) in patients treated with gross-total resection (GTR), subtotal resection (STR), GTR + radiation (RT), or STR + RT. B, Kaplan-Meier survival curves depicting the differences in survival for ependymoma NOS in patients treated with GTR, STR, GTR + RT, or STR + RT. C, Kaplan-Meier survival curves depicting the differences in survival for anaplastic ependymoma, NOS in patients treated with GTR, STR, GTR + RT, or STR + RT. (Central Brain Tumor Registry of the United States 2000-2016).
Multivariable Cox proportional hazard regression models were used to further examine the association of clinical and demographic characteristics with overall survival in anaplastic ependymoma and ependymoma, NOS (Table 2). Among the total sample, those diagnosed with anaplastic ependymoma had roughly 3 times the risk of death compared to individuals diagnosed with ependymoma, NOS (hazard ratio [HR]: 2.93 [95% CI, 2.19-3.92]). African Americans diagnosed with either ependymoma had a 78% increased risk of death compared to whites (HR: 1.78 [95% CI, 1.30-2.44]) and a 91% increase in mortality for ependymoma, NOS (HR: 1.91 [95% CI, 1.31-2.79]). However, the risk of death for API vs white, or Hispanic vs non-Hispanic individuals was not significantly different. Adults (age 40+ years) had almost twice the risk of death compared to children (0-14 years) (HR: 1.97 [95% CI, 1.45-2.66]) for all ependymoma and more than 5 times the mortality in anaplastic ependymoma (HR: 5.05 [95% CI, 3.08-8.28]). Overall, ependymomas in the supratentorial region were associated with a risk of mortality 2.5 times greater than spinal cord ependymomas (HR: 2.50 [95% CI, 1.77-3.53]) (see Table 2). Likewise, ependymomas in the infratentorial region were associated with a risk of mortality 2.76 times greater than spinal cord ependymomas (HR: 2.76 [95% CI, 2.07-3.68]) (see Table 2). In all ependymomas, individuals undergoing STR had a risk of mortality 1.81 times greater than individuals undergoing GTR (HR: 1.81 [95% CI, 1.33-2.46]). Overall, combined STR and RT did not have significantly higher mortality compared to GTR alone (see Table 2). There was a statistically significant reduced risk of death in individuals treated with GTR and RT vs GTR alone for anaplastic ependymoma (HR: 0.50 [95% CI, 0.30-0.85]) (see Table 2).
Table 2.
Multivariable Cox Proportional Hazards Model Results for Anaplastic Ependymoma and Ependymoma, Not Otherwise Specified; Surveillance Epidemiology and End Results 2000-2016
| Overall | Ependymoma, NOS | Anaplastic ependymoma | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Factor | HR | 95% CI | P | HR | 95% CI | P | HR | 95% CI | P |
| Ependymoma, anaplastic vs ependymoma, NOS | 2.93 | 2.19-3.92 | < .001a | ||||||
| African American vs white | 1.78 | 1.30-2.44 | < .001a | 1.91 | 1.31-2.79 | .001a | 1.58 | 0.90-2.76 | .11 |
| API vs white | 1.09 | 0.70-1.70 | .70 | 0.99 | 0.53-1.84 | .97 | 1.58 | 0.82-3.04 | .17 |
| Hispanic vs non-Hispanic | 1.29 | 0.98-1.70 | .07 | 1.26 | 0.89-1.78 | .19 | 1.33 | 0.84-2.12 | .23 |
| Male vs female | 1.07 | 0.86-1.32 | .54 | 1.02 | 0.78-1.32 | .90 | 1.37 | 0.94-2.00 | .10 |
| AYA vs child | 0.93 | 0.66-1.30 | .67 | 0.67 | 0.43-1.03 | .07 | 1.31 | 0.77-2.25 | .32 |
| Adult vs child | 1.97 | 1.45-2.66 | < .001a | 1.24 | 0.85-1.82 | .26 | 5.05 | 3.08-8.28 | < .001a |
| Supratentorial vs spinal | 2.50 | 1.77-3.53 | < .001a | 2.43 | 1.56-3.78 | < .001a | 2.36 | 1.11-5.00 | .03a |
| Infratentorial vs spinal | 2.76 | 2.07-3.68 | < .001a | 2.51 | 1.82-3.46 | < .001a | 3.48 | 1.55-7.79 | .002a |
| STR only vs GTR only | 1.81 | 1.33-2.46 | < .001a | 1.86 | 1.32-2.63 | < .001a | 1.42 | 0.70-2.88 | .33 |
| GTR and RT vs GTR only | 0.78 | 0.57-1.07 | .12 | 0.87 | 0.58-1.30 | .50 | 0.50 | 0.30-0.85 | .01a |
| STR and RT vs GTR only | 1.38 | 1.01-1.89 | .046a | 1.38 | 0.94-2.02 | .10 | 0.97 | 0.54-1.73 | .91 |
Abbreviations: AYA, adolescent and young adult; GTR, gross total resection; HR, hazard ratio; NOS, not otherwise specified; RT, radiation therapy; STR, subtotal resection.
aIndicates significant P value.
Discussion
In this report, we provide an in-depth examination of 2 ependymoma histologies, anaplastic ependymoma and ependymoma, NOS incidence, and survival trends in the United States over a 15-year period from 2000 to 2016. We found that the incidence of anaplastic ependymoma was highest in age groups 0 to 4 years followed by age 5 to 9 years and 10 to 14 years respectively before plateauing in adults and subsequently declining in individuals older than 85 years. The incidence of ependymoma, NOS was highest in ages 0 to 4 years in the pediatric population, peaked in the 45- to 69-year age group, and declined again in octogenarians. In pediatric populations, the incidence of anaplastic ependymoma and ependymoma, NOS were similar in ages 0 to 4 years only, with the rest of the pediatric age groups showing significantly higher ependymoma, NOS incidence compared to anaplastic ependymoma. These trends are consistent with prior reports of ependymoma incidence in the United States, demonstrating stable distributions over an extended time period.3,5 Ependymoma incidence trends in this study are similar to other CNS tumors such as oligodendroglioma and anaplastic oligodendroglioma, but are in contrast to the increased incidence of meningioma in the elderly.16,17 These results, and those of other studies, reinforce the importance of studies contributing to the descriptive epidemiology of CNS tumors.4,18
Interestingly, the incidence of anaplastic ependymoma in ages 0 to 4 and 5 to 9 years significantly increased over time in contrast to a significant decrease in ependymoma, NOS in these age groups over the same time period. It is unclear to what these differences are attributable, but the improved understanding of the molecular heterogeneity of ependymoma highlights the age-specific distributions of YAP1-fusion–positive and RELA-fusion–positive ependymoma, for example.19 However, the etiology is likely multifactorial, including intrinsic tumor behavior, environmental, and classification/pathology reporting changes over time.20 There was no significant difference in anaplastic ependymoma incidence between male and female patients. In contrast, the incidence of ependymoma, NOS was significantly higher in male compared to female patients. The male predominance may suggest that sex modifies molecular subtype-specific tumor behavior, likely relating to sex-related alterations in tumorigenesis and neuronal progenitor population responses to sex hormones.21–23 The lack of sex-specific IRs in anaplastic ependymoma may suggest the differences related to anaplastic-type molecular signatures. Nonetheless, both findings provoke questions and research into molecular-based tumor behavior and response to treatment. In both ependymoma, NOS and anaplastic ependymoma for male and female patients, the incidence of each significantly increased from 2000 until 2004 to 2005 and then significantly decreased until 2016.
Like the incidence of race in many other primary CNS tumor histologies, white populations had the highest incidence of anaplastic ependymoma and ependymoma, NOS, with the exception of API populations, whose incidence of anaplastic ependymoma was not significantly different from white populations. In contrast, non-Hispanic populations had a significantly higher incidence of ependymoma, NOS compared to Hispanics, though there was no difference in anaplastic ependymoma. White populations experienced significant increases in the incidence of ependymoma, NOS from 2000 to 2004 and then significant declines from 2004 to the end of the study period. In contrast, the incidence of anaplastic ependymoma increased in white, African American, and Hispanic populations from 2000 to 2016. As previously mentioned, although these incidence trends are similar to other CNS tumors and likely reflect true predominance in these populations, IR may be biased toward higher reporting in non-Hispanic whites given previous reports of their greater access to care and earlier diagnoses.24
An interesting pattern was observed for the location distribution of anaplastic ependymoma compared to ependymoma, NOS. In particular, the incidence of anaplastic ependymoma was highest in supratentorial locations, followed by infratentorial locations, with the lowest incidence noted in the spinal cord. However, the incidence distribution of location for ependymoma, NOS was opposite to anaplastic ependymoma, with the highest incidence in the spinal cord, followed by infratentorial, then supratentorial lesions. Incidence of anaplastic ependymomas in all 3 locations increased significantly from 2000 to 2016, whereas spinal cord and supratentorial ependymoma, NOS incidence increased from 2000 to 2003 and then declined thereafter. The distinctions between CNS compartment incidences is a phenomenon relating to molecular subgrouping. Molecular subtype research has revealed several patterns of ependymoma tumorigenesis specific to each compartment.21 These include posterior fossa group A and B ependymomas,25 supratentorial RELA- vs YAPI-fusion–positive ependymomas,26 and spinal classic vs myxopapillary ependymomas, all with distinct molecular profiles and prognoses.19 The incidence trends described here help reinforce the necessity to better characterize and provide trend location with subtype-specific tumor behaviors over time.
Overall survival was examined for different groups of individuals both for anaplastic ependymoma and ependymoma, NOS. Not surprisingly, individuals with anaplastic ependymoma had a 2 times higher mortality risk compared to ependymoma, NOS. In both all ependymoma and ependymoma, NOS, African American populations had a 78% and 91% increased risk of death, respectively, compared to white populations. In anaplastic ependymoma, adults (age 40+ years) had a 5 times greater risk of death compared to children (age 0-14 years). In contrast, young adults (age 15-39 years) trended as having a lower risk of death compared to children with ependymoma, NOS, demonstrating clear trend differences in age-related mortality when comparing anaplastic and NOS ependymoma. In terms of CNS compartment, supratentorial ependymomas were associated with a risk of mortality 2.5 times greater than spinal cord ependymomas, and infratentorial ependymomas were associated with a risk of mortality 2.76 times greater than spinal ependymomas. This mortality trend held true both for NOS and anaplastic ependymoma (see Table 2). Although several studies have demonstrated sex, socioeconomic status, ethnicity, and CNS location survival differences in glioma,27 here we report the most recent, comprehensive description of differences in currently available US population-based data for ependymoma. In fact, other recent analyses of CNS tumors show increased glioma IRs and lower survival rates in non-Hispanic whites, which is distinct from the poorer survival we see in African American populations with ependymoma.24 The discrepancies in survival noted previously, and highlighted in this study, may speak to important health disparities that remain unresolved.
In multivariable Cox proportional hazard regression models for ependymoma, NOS, individuals undergoing STR had a risk of mortality 1.86 times greater than individuals undergoing GTR, indicative of the importance of surgical treatment. These patients still had a 1.38 times greater risk of mortality with adjuvant RT (STR + RT) compared to GTR alone. For anaplastic ependymoma, there was no statistically significant mortality advantage in individuals treated with GTR and RT vs GTR alone. These findings reaffirm results from other studies highlighting the importance of GTR.11,21,28 However, it is interesting to note in this study that RT with GTR did not convey an overall survival benefit compared to GTR alone, suggesting that a survival benefit to RT is significant in the setting of unresected or residual tumor. This result could alternatively indicate that the benefit of RT may depend on distinct molecular profiles and their specific radiosensitivity. On the other hand, this result could be confounded by limitations in available sample size, thereby prohibiting this study from including treatment in a multivariable analysis. Further research is warranted to determine the advantage of adjuvant RT after GTR to develop treatment algorithms that provide maximal benefit with minimal risk to these patient populations.
Strengths and Limitations
In this study, we recognize the limitation of classifying ependymomas as either NOS or anaplastic, given that there are multiple distinct subtypes. However, given the rarity of these tumors, and, therefore the small sample sizes obtained in subtype-specific analyses, examining the incidence of anaplastic vs other types of ependymoma is beneficial in analyzing overall changes/trends that will direct further in-depth analysis. Additionally, it is clear that ependymoma molecular signatures have a significant impact on malignant behavior and response to treatment, limiting the utility of histopathological classification risk stratification. Future trials26,29 and database design require inclusion of molecular data to improve tumor-specific treatment and prognostication. In addition, there are limitations inherent to the collection of data from national databases such as CBRTUS, NPCR, and SEER registries. Though CBTRUS, in collaboration with the CDC and NCI, represents the largest and most up-to-date population-based registry focused exclusively on CNS tumors, and represents cases collected from the entire US population, there is no central pathology review of cases within the US cancer registry system. Thus, histology code assignment at case registration is based on histology information contained in the patient’s medical record. As a result, despite the most recent revision in 2016 of the World Health Organization Classification of Tumors of the Central Nervous System, tumors reported beforehand may have been diagnosed using any of the prior classifications because of the variation in adoption of new standards among participating institutions. Histologies are reflective, then, of the prevailing criteria for a histology at the time of registration and not the most recent classifications. Incidence trends may also be affected by increased case ascertainment and identification, particularly among primary brain and CNS tumors, over time.
This study reports differences in incidence and survival for anaplastic and ependymoma, NOS based on age, race/ethnicity, sex, histology, and location within the nervous system. African American populations experience higher mortality rates despite lower incidence compared to white populations. Additionally, STR and RT still had higher mortality rates compared to GTR alone both in anaplastic and NOS ependymoma. These results highlight the need for further study regarding treatment patterns and racial disparities in neuro-oncological care.
Supplementary Material
Funding
CBTRUS was supported by the Centers for Disease Control and Prevention (CDC) [under contract No. 75D30119C06056], the American Brain Tumor Association, The Sontag Foundation, Novocure, the Musella Foundation, National Brain Tumor Society, the Zelda Dorin Tetenbaum Memorial Fund, as well as private and in-kind donations. Contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC.
Conflict of interest statement. None declared.
References
- 1. Taylor MD, Poppleton H, Fuller C, et al. Radial glia cells are candidate stem cells of ependymoma. Cancer Cell. 2005;8(4):323–335. [DOI] [PubMed] [Google Scholar]
- 2. Armstrong T, Yuan Y, Wu J, et al. Objective response and clinical benefit in recurrent ependymoma in adults: final report of CERN 08-02: a phase II study of dose-dense temozolomide and lapatinib. Neuro Oncol. 2018;20(Suppl 6):vi241. [Google Scholar]
- 3. Ostrom QT, Cioffi G, Gittleman H, et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2012-2016. Neuro Oncol. 2019;21(Suppl 5):v1–v100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gittleman HR, Ostrom QT, Rouse CD, et al. Trends in central nervous system tumor incidence relative to other common cancers in adults, adolescents, and children in the United States, 2000 to 2010. Cancer. 2015;121(1):102–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Villano JL, Parker CK, Dolecek TA. Descriptive epidemiology of ependymal tumours in the United States. Br J Cancer. 2013;108(11):2367–2371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. McGuire CS, Sainani KL, Fisher PG. Both location and age predict survival in ependymoma: a SEER study. Pediatr Blood Cancer. 2009;52(1):65–69. [DOI] [PubMed] [Google Scholar]
- 7. McGuire CS, Sainani KL, Fisher PG. Incidence patterns for ependymoma: a Surveillance, Epidemiology, and End Results study. J Neurosurg. 2009;110(4):725–729. [DOI] [PubMed] [Google Scholar]
- 8. Amirian ES, Armstrong TS, Gilbert MR, Scheurer ME.. Predictors of survival among older adults with ependymoma. J Neurooncol. 2012;107(1):183–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Akyurek S, Chang EL, Yu TK, et al. Spinal myxopapillary ependymoma outcomes in patients treated with surgery and radiotherapy at M.D. Anderson Cancer Center. J Neurooncol. 2006;80(2):177–183. [DOI] [PubMed] [Google Scholar]
- 10. Agbahiwe HC, Wharam M, Batra S, Cohen K, Terezakis SA.. Management of pediatric myxopapillary ependymoma: the role of adjuvant radiation. Int J Radiat Oncol Biol Phys. 2013;85(2):421–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Khalid SI, Adogwa O, Kelly R, et al. Adult spinal ependymomas: an epidemiologic study. World Neurosurg. 2018;111:e53–e61. [DOI] [PubMed] [Google Scholar]
- 12. Khalid SI, Kelly R, Adogwa O, et al. Pediatric spinal ependymomas: an epidemiologic study. World Neurosurg. 2018;115:e119–e128. [DOI] [PubMed] [Google Scholar]
- 13. Merchant TE, Kiehna EN, Thompson SJ, Heideman R, Sanford RA, Kun LE. Pediatric low-grade and ependymal spinal cord tumors. Pediatr Neurosurg. 2000;32(1):30–36. [DOI] [PubMed] [Google Scholar]
- 14. Nuño M, Yu JJ, Varshneya K, et al. Treatment and survival of supratentorial and posterior fossa ependymomas in adults. J Clin Neurosci. 2016;28:24–30. [DOI] [PubMed] [Google Scholar]
- 15. Safaee M, Oh MC, Kim JM, et al. Histologic grade and extent of resection are associated with survival in pediatric spinal cord ependymomas. Childs Nerv Syst. 2013;29(11):2057–2064. [DOI] [PubMed] [Google Scholar]
- 16. Achey RL, Gittleman H, Schroer J, Khanna V, Kruchko C, Barnholtz-Sloan JS. Nonmalignant and malignant meningioma incidence and survival in the elderly, 2005-2015, using the Central Brain Tumor Registry of the United States. Neuro Oncol. 2019;21(3): 380–391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. .Achey RL, Khanna V, Ostrom QT, Kruchko C, Barnholtz-Sloan JS.. Incidence and survival trends in oligodendrogliomas and anaplastic oligodendrogliomas in the United States from 2000 to 2013: a CBTRUS report. J Neurooncol. 2017;133(1):17–25. [DOI] [PubMed] [Google Scholar]
- 18. Ostrom QT, Gittleman H, de Blank PM, et al. American Brain Tumor Association adolescent and young adult primary brain and central nervous system tumors diagnosed in the United States in 2008-2012. Neuro Oncol. 2016;18(Suppl 1):i1–i50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Pajtler KW, Witt H, Sill M, et al. Molecular classification of ependymal tumors across all CNS compartments, histopathological grades, and age groups. Cancer Cell. 2015;27(5):728–743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Arora RS, Alston RD, Eden TO, Estlin EJ, Moran A, Birch JM.. Age-incidence patterns of primary CNS tumors in children, adolescents, and adults in England. Neuro Oncol. 2009;11(4):403–413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Khatua S, Ramaswamy V, Bouffet E. Current therapy and the evolving molecular landscape of paediatric ependymoma. Eur J Cancer. 2017;70:34–41. [DOI] [PubMed] [Google Scholar]
- 22. Mack SC, Taylor MD. Put away your microscopes: the ependymoma molecular era has begun. Curr Opin Oncol. 2017;29(6):443–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sun T, Plutynski A, Ward S, Rubin JB.. An integrative view on sex differences in brain tumors. Cell Mol Life Sci. 2015;72(17):3323–3342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Ostrom QT, Cote DJ, Ascha M, Kruchko C, Barnholtz-Sloan JS.. Adult glioma incidence and survival by race or ethnicity in the United States from 2000 to 2014. JAMA Oncol. 2018;4(9):1254–1262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Witt H, Mack SC, Ryzhova M, et al. Delineation of two clinically and molecularly distinct subgroups of posterior fossa ependymoma. Cancer Cell. 2011;20(2):143–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Parker M, Mohankumar KM, Punchihewa C, et al. C11orf95-RELA fusions drive oncogenic NF-κB signalling in ependymoma. Nature. 2014;506(7489):451–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Tseng MY, Tseng JH, Merchant E. Comparison of effects of socioeconomic and geographic variations on survival for adults and children with glioma. J Neurosurg. 2006;105(4 Suppl):297–305. [DOI] [PubMed] [Google Scholar]
- 28. Merchant TE, Li C, Xiong X, Kun LE, Boop FA, Sanford RA.. Conformal radiotherapy after surgery for paediatric ependymoma: a prospective study. Lancet Oncol. 2009;10(3):258–266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Godfraind C. Classification and controversies in pathology of ependymomas. Childs Nerv Syst. 2009;25(10):1185–1193. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.