Incidence and survival of primary central nervous system tumors diagnosed in 4 Canadian provinces from 2010 to 2015

Abstract

Background

The Brain Tumor Registry of Canada was established in 2016 to enhance infrastructure for surveillance and clinical research on Central Nervous System (CNS) tumors. We present information on primary CNS tumors diagnosed among residents of Canada from 2010 to 2015.

Methods

Data from 4 provincial cancer registries were analyzed representing approximately 67% of the Canadian population. Age-standardized incidence rates (ASIR) and 95% confidence intervals (CI) were calculated using the 2011 Canadian population age distribution. Net survival was estimated using the Pohar-Perme method.

Results

A total of 31 644 primary tumors were identified for an ASIR of 22.8 per 100 000 person-years. Nonmalignant tumors made up 47.1% of all classified tumors, with mixed behaviors present in over half of histology groupings. Unclassified were 19.5% of all tumors. The most common histological subtypes are meningiomas (ASIR = 5.5 per 100 000 person-years); followed by glioblastomas (ASIR 4.0 per 100 000 person-years). The overall 5-year net survival rate for CNS tumors was 65.5%; females 70.2% and males 60.4%. GBMs continue to be the most lethal CNS tumors for all sex and age groups.

Conclusions

The low annual frequency of most CNS tumor subtypes emphasizes the value of population-based data on all primary CNS tumors diagnosed among Canadians. The large number of histological categories including mixed behaviors and the proportion of unclassified tumors emphasizes the need for complete reporting. Variation in incidence and survival across histological groups by sex and age highlights the need for comprehensive and histology-specific reporting. These data can be used to better inform research and health system planning.

The Brain Tumour Registry of Canada project was established in 2016 with the aim of enhancing infrastructure for surveillance and clinical research to improve health outcomes for central nervous system (CNS) tumor patients in Canada.¹ Until recently, population-level data on nonmalignant CNS tumors among Canadians was imputed using data from the United States.² However, this was deemed insufficient by a special committee of the Brain Tumour Foundation of Canada, as the resulting surveillance estimates did not accurately capture the distribution of CNS patients in Canada. In response, the Canadian House of Commons Bill M235 was passed in February 2007, mandating the collection of data on nonmalignant tumors in provincial/territorial cancer registries, and requesting the creation of national guidelines for the surveillance of CNS tumors.³ Because nonmalignant CNS tumors are diagnosed using different methods from malignant tumors, cancer registry workflows are impacted. As such, information on these tumors is currently incomplete in the Canadian Cancer Registry system and case capture varies by province, demographics, tumor characteristics, and year of diagnosis.⁴

Overall, CNS tumors are rare, accounting for 1%–2% of all tumors included in the cancer registries.⁵ However, CNS tumors are a heterogeneous group of tumors, with over 120 tumor subtypes in the WHO classification schema.⁶ This results in a challenging set of tumors to report effectively. In general, brain cancers are recognized as having a very poor prognosis; less well-recognized is that prognosis of CNS tumors varies widely by histological classification with a few having a much better prognosis. Routine surveillance reports combine all CNS tumors to provide overall estimates of their annual frequency. This is of limited use to the neuro-oncology community, for which comprehensive information on the incidence and survival of histological subtypes of all primary CNS tumors is crucial to inform research and health system planning. This report incorporates information on histological subgroups of primary CNS tumors obtained from 4 provincial cancer registries with a focus on the value added by incorporating the nonmalignant tumors.

Methods

Data Sources and Management

Each provincial registry collects patient data under the authority of their respective legislation: The Cancer Act and the Provincial Personal Health Information Protection Act.⁵ Across provinces there is some variation in the data sources used to identify new cases (Jiang Y., Benign brain tumor surveillance project [unpublished]. Public Health Agency of Canada, Chronic Disease Surveillance Division; 2009 (2009)). Four provinces were engaged in this effort, reflecting 67% of the Canadian population: Manitoba, with a history of collecting information on all primary brain tumors; Ontario, which incorporated nonmalignant brain tumor information from 2010 forward; Alberta and British Columbia, which used the discharge abstract database as a new case-reporting source to supplement routine case ascertainment beginning in 2010.⁷ Patients diagnosed with primary CNS tumors registered in the 4 provincial cancer registries between 2010 and 2015 were identified. Files from Ontario (n = 21 261), Manitoba (n = 1573), Alberta (n = 4167), and British Columbia (n = 6359) were combined. An incidence analytic file (n = 31 644) was created including CNS tumor codes (Tumor Classification below) and excluding any patients diagnosed after 2015 (Supplementary Figure S1). A separate survival analytic file was created (n = 31 167) with 2 additional exclusions: cases coded as autopsy or death certificate only and patients with the same date of diagnosis and date of death (Supplementary Figure S2).

Denominators for all incident rate calculations were obtained from Statistics Canada.⁸ Post-census and inter-census estimates were used for each year, by age and sex. Life tables from the Canadian Socio-Economic Information Management System (CANSIM) were used from 2010 to 2016 for net survival analysis and matched to the data by age, sex, year of diagnosis, and province of diagnosis.⁹ Data files were combined by project analysts at the University of Alberta. Data analysis was conducted using SAS version 9.4. Figures were generated using R version 4.0.3.

Data in tables are suppressed to preserve the privacy of patients, in accordance with reporting practices of Statistics Canada. Estimates were suppressed if there were <5 cases diagnosed over the 6-year period. Net survival estimates were censored when the total number of cases was <50, or <5 deaths in that category. Furthermore, cells that could result in residual disclosure were suppressed. Residual disclosure occurs when cross-referencing data across tables leads to cell counts that are less than 5. All counts underwent random rounding procedures, rounding estimates to the nearest multiple of 5.

Ethics Approvals and Data Sharing Agreements

Ethics approval for this project was first granted by the Health Research Ethics Board of Alberta (HREBA) Cancer Committee, and then subsequently by ethics committees in British Columbia, Manitoba, and Ontario. Data-sharing agreements were established with each participating provincial cancer registry.

Tumor Classification

Disease groups presented in this report incorporate histological features and behavior. The classification system used by cancer registries in Canada is the International Classification of Diseases for Oncology, 3rd edition, or ICDO-3,¹⁰ which is aligned with 2007 WHO CNS 4th edition classification.¹¹ This multiaxial classification system assigns alphanumeric codes for the anatomical site of the tumor (topography) and numeric codes for the histology and behavior. Primary CNS tumors were defined as those occurring at the following ICDO-3 sites: C70.0–C70.9, C71.0–71.9, C72.0–C72.9, C75.1–C75.3, and C300 (limited to histology codes 9522–9523). Histology codes were grouped into categories based on classifications used by the Central Brain Tumor Registry of the United States (CBTRUS).¹² Tumors located in the CNS but with a histology code not included in the CBTRUS scheme were called “not classified in CBTRUS.” Nonmalignant, borderline and malignant behavior cases are included in this report within their appropriate histology grouping. For further details by tumor location and behavior groupings please refer to https://braintumourregistry.ca. At present, data on the molecular features of these tumors are not routinely collected by provincial cancer registries.

Data Analysis

Estimates produced for this report include counts, averages, proportions, and rates. Estimates are stratified by histology, sex, and age group and are reported to 1 significant digit. Age was categorized to be consistent with recent cancer surveillance reports¹³: children were 0–14 years of age, adolescents and young adults (AYA) were 15–39 years of age and adults were over age 40. Adults were further categorized into the middle (40–59 years) and older adults (60+ years). All incidence rates were age-standardized using the 2011 Canadian population as a reference.⁸ Incidence rates and 95% confidence intervals (CI) are reported for major (denoted in bold text) and specific subtype categories of tumors. Four histological subtypes had fewer than 10 cases per year (pineal, other neuroepithelial, primary melanocytic, other hematopoietic) making those estimates very unstable by sex and age.

Patient follow-up time for survival was defined as the time between the date of diagnosis and the date of death or the follow-up cutoff date defined by each province. Cutoff dates were defined by each province and rather than shorten some follow-up dates we allowed them to vary by province: Ontario was December 31, 2017; Manitoba was December 31, 2017; British Columbia was July 18, 2018, and Alberta was September 30, 2018. 1 and 5-year net survival estimates and associated CI were calculated using the cohort approach and the Pohar-Perme method. Net survival is an estimate of the probability of surviving cancer, adjusting for deaths from other causes. By incorporating the survival experience of the general population in the calculation, this measure also accounts for changing trends in causes of death in the underlying population.¹⁴ The 2016 province-specific life table was used until the end of the follow-up period.⁹ Net survival is only reported for specific subtypes, as within-group heterogeneity in survival reduces the utility of estimates from broader histology categories.

Results

Incidence

An overall incidence rate of 22.8 per 100 000 person-years (95%CI = 22.2–23.4) was estimated. Within tumor types, incidence rates are generally higher for males than females (Table 1). The preponderance of female meningiomas is reflected in total rates being lower in males (ASIR_males = 22.3 per 100 000 person-years, 95%CI = 21.4–23.2) than females (ASIR_females = 23.3per 100 000 person-years, 95%CI = 22.5–24.2).

Table 1.

All Primary Brain and Other Central Nervous System (CNS) Tumors By Histology Group and Sex: Average Number of Cases Per Year(), Percent Malignant (% M), Average Annual Age-Standardized Incidence Rates^a (ASIR) per 100 000 and 95% Confidence Intervals (CI) From 4 Provinces (British Columbia, Alberta, Manitoba, and Ontario) 2010–2015

	Total				Male				Female
Histology Group (Major/ Specific)b		% M	ASIR	95% CI		% M	ASIR	95% CI		% M	ASIR	95% CI
Tumors of neuroepithelial tissue	1689	92.9	7.3	7.0–7.6	977	93.1	8.7	8.2–9.3	712	92.7	6.0	5.5–6.4
Pilocytic astrocytoma	60	100.0	0.3	0.2–0.3	30	100.0	0.3	0.2–0.3	30	100.0	0.3	0.2–0.4
Diffuse astrocytoma	84	100.0	0.4	0.3–0.4	49	100.0	0.4	0.3–0.6	34	100.0	0.3	0.2–0.4
Anaplastic astrocytoma	72	100.0	0.3	0.2–0.4	39	100.0	0.3	0.2–0.5	33	100.0	0.3	0.2–0.4
Unique astrocytoma variants	16	70.4	0.1	0.04–0.1	10	68.9	0.1	0.0–0.1	6	73.0	0.1	0.0–0.1
Glioblastoma	912	100.0	4.0	3.7–4.2	540	100.0	4.9	4.5–5.3	372	100.0	3.1	2.8–3.4
Oligodendroglioma	67	100.0	0.3	0.2–0.4	40	100.0	0.3	0.2–0.5	27	100.0	0.2	0.1–0.3
Anaplastic oligodendroglioma	56	100.0	0.2	0.2–0.3	32	100.0	0.3	0.2–0.4	23	100.0	0.2	0.1–0.3
Choroid plexus tumors	11	10.9	0.05	0.02–0.1	5	11.1	0.04	0.0–0.1	6	10.8	0.1	0.0–0.1
Oligoastrocytic tumors	67	100.0	0.3	0.2–0.4	38	100.0	0.3	0.2–0.4	29	100.0	0.2	0.2–0.3
Ependymal tumors	93	58.7	0.4	0.3–0.5	51	57.1	0.4	0.3–0.6	42	60.6	0.4	0.3–0.5
Glioma malignant, NOSc	117	100.0	0.5	0.4–0.6	62	100.0	0.6	0.4–0.7	55	100.0	0.5	0.3–0.6
Neuronal & mixed neuronal-glial tumors	76	17.4	0.3	0.3–0.4	44	18.5	0.4	0.3–0.5	32	15.8	0.3	0.2–0.4
Tumors of the pineal region	9	67.3	0.04	0.01–0.1	5	64.3	0.04	0.0–0.1	4	70.8	0.03	0.0–0.1
Embryonal tumors	50	97.7	0.2	0.2–0.3	31	98.4	0.3	0.2–0.4	19	96.6	0.2	0.1–0.2
Other neuroepithelial tumorsd	1	80.0	0.004	0.0–0.01	1	66.7	0.004	0.0–0.02	.	.	0.003	0.0–0.01
Tumors of cranial & spinal nerves	408	0.9	1.8	1.6–1.9	200	0.8	1.8	1.5–2.0	208	1.0	1.8	1.5–2.0
Tumors of the Meninges	1347	3.0	5.8	5.5–6.2	420	4.7	3.9	3.5–4.2	927	2.3	7.7	7.2–8.2
Meningioma	1262	2.1	5.5	5.2–5.8	378	3.0	3.5	3.1–3.9	884	1.7	7.3	6.8–7.8
Mesenchymal tumors	28	35.5	0.1	0.1–0.2	13	42.5	0.1	0.1–0.2	14	29.1	0.1	0.1–0.2
Primary Melanocytic lesions	4	63.6	0.02	0.0–0.03	2	66.7	0.02	0.0–0.04	2	60.0	0.01	0.0–0.04
Other, related to the meninges	54	3.7	0.2	0.2–0.3	27	4.3	0.2	0.2–0.3	27	3.1	0.2	0.1–0.3
Lymphomas & hematopoietic neoplasms	105	99.8	0.5	0.3–0.5	59	100.0	0.5	0.4–0.7	46	99.6	0.4	0.3–0.5
Lymphoma	103	100.0	0.4	0.4–0.5	58	100.0	0.5	0.4–0.7	44	100.0	0.4	0.3–0.5
Other hematopoietic neoplasms	2	92.9	0.01	0.0–0.02	1	100.0	0.01	0.0–0.02	2	90.0	0.01	0.0–0.03
Germ cell tumors, cysts, & heterotopias	26	62.7	0.1	0.1–0.2	19	71.4	0.2	0.1–0.2	8	41.3	0.1	0.0–0.1
Tumors of the sellar region	647	0.4	2.8	2.6–3.0	329	0.6	3.0	2.6–3.3	318	0.3	2.7	2.4–3.0
Unclassified tumors	1028	13.7	4.4	4.2–4.7	437	14.9	4.2	3.8–4.6	591	12.8	4.7	4.3–5.1
Not classified in CBTRUS	24	25.5	0.1	0.1–0.1	13	28.6	0.1	0.1–0.2	11	21.9	0.1	0.0–0.2
Total	5274	35.7	22.8	22.2–23.4	2454	43.7	22.3	21.4–23.2	2820	28.8	23.3	22.5–24.2

^aRates are age-standardized to the 2011 Canadian standard population.

^bDefined as per the Central Brain Tumor Registry of the United States.¹³

^cNOS = not otherwise specified.

^dEstimates based on fewer than 5 observed cases over the 6-year period are suppressed.

The most common histologies varied considerably by age at diagnosis (Table 2). In children the most common CNS tumor types are pilocytic astrocytoma (ASIR = 0.8 per 100 000 person-years, 95%CI = 0.6–1.1) and embryonal tumors (ASIR = 0.8 per 100 000 person-years, 95%CI = 0.6–1.1); in AYA the most common are in the sellar region (ASIR = 1.8 per 100 000 person-years, 95%CI = 1.5–2.1) and meningiomas (ASIR = 1.1 per 100 000 person-years, 95%CI = 0.9–1.3); while in adults the most common are meningiomas (ASIR = 10.2 per 100 000 person-years, 95%CI = 9.6–10.8) and glioblastomas (ASIR = 7.5 per 100 000 person-years, 95%CI = 7.0–8.0). The incidence rates increased with age. In AYA (ASIR = 9.6 per 100, 000 person-years, 95%CI = 8.9–10.3) are almost double that of children (5.5 per 100 000 person-years, 95%CI = 4.8–6.3), while adult incidence rates (ASIR = 37.4 per 100 000 person-years, 95%CI = 36.3–38.51) are almost 7 times higher than in children.

Table 2.

All Primary Brain and Other Central Nervous System (CNS) Tumors By Age Group (Children, Adolescents/Young Adults (AYA) and Adults) and Histology Groups: Average Number of Cases Per Year (), Percent Malignant (% M), Average Annual Age-standardized Incidence Rates^a (ASIR) Per 100 000 and 95% Confidence Intervals (CI) From 4 Provinces (British Columbia, Alberta, Manitoba, and Ontario) 2010–2015

	Children (0–14) Yrs				AYA (15–39) Yrs				Adults (40+) Yrs
Histology Group (Major/Specific)b		% M	ASIR	95% CI		% M	ASIR	95% CI		% M	ASIR	95% CI
Tumors of neuroepithelial tissue	149	85.2	3.9	3.3–4.5	283	81.9	3.6	3.2–4.0	1257	96.3	10.9	10.3–11.5
Pilocytic astrocytoma	33	100.0	0.8	0.6–1.1	18	100.0	0.2	0.1–0.3	9	100.0	0.1	0.02–0.1
Diffuse astrocytoma	9	100.0	0.2	0.1–0.4	24	100.0	0.3	0.2–0.4	51	100.0	0.4	0.3–0.6
Anaplastic astrocytoma	2	100.0	0.1	0.0–0.1	22	100.0	0.3	0.2–0.4	47	100.0	0.4	0.3–0.5
Unique astrocytoma variants	3	40.0	0.1	0.0–0.2	7	62.5	0.1	0.02–0.2	6	94.7	0.1	0.01–0.1
Glioblastoma	7	100.0	0.2	0.1–0.3	44	100.0	0.6	0.4–0.7	861	100.0	7.5	7.0–8.0
Oligodendrogliomad	.	.	.	.-.	27	100.0	0.3	0.2–0.5	39	100.0	0.3	0.2–0.5
Anaplastic oligodendrogliomad	.	.	.	.-.	13	100.0	0.2	0.1–0.3	43	100.0	0.4	0.3–0.5
Choroid plexus tumors	5	19.4	0.1	0.02–0.3	2	7.7	0.03	0.0–0.1	3		0.03	0.0–0.1
Oligoastrocytic tumorsd	.	.	.	.-.	23	100.0	0.3	0.2–0.4	44	100.0	0.4	0.3–0.5
Ependymal tumors	14	90.5	0.4	0.2–0.6	26	58.4	0.3	0.2–0.5	54	50.5	0.5	0.3–0.6
Glioma malignant, NOSc	26	100.0	0.7	0.4–0.9	25	100.0	0.3	0.2–0.4	66	100.0	0.6	0.4–0.7
Neuronal & mixed neuronal-glial tumors	15	3.4	0.4	0.2–0.6	38	7.9	0.5	0.3–0.6	23	41.7	0.2	0.1–0.3
Tumors of the pineal region	2	100.0	0.1	0.0–0.1	2	71.4	0.0	0.0–0.1	4	45.8	0.03	0.0–0.1
Embryonal tumors	32	99.0	0.8	0.6–1.1	12	97.1	0.1	0.1–0.2	6	92.1	0.1	0.01–0.1
Other neuroepithelial tumorsd	1	66.7	0.01	0.0–0.1	.	.	.	.-.	.	.	.	.-.
Tumors of cranial & spinal nerves	9	5.9	0.2	0.1–0.4	84	1.6	1.1	0.8–1.3	315	0.6	2.7	2.4–3.1
Tumors of the Meninges	6	48.6	0.2	0.03–0.3	109	3.7	1.4	1.1–1.6	1232	2.7	10.7	10.1–11.3
Meningioma	3	12.5	0.1	0.0–0.2	88	2.1	1.1	0.9–1.3	1171	2.1	10.2	9.6–10.8
Mesenchymal tumors	2	83.3	0.1	0.0–0.1	4	47.6	0.04	0.0–0.1	22	29.3	0.2	0.1–0.3
Primary Melanocytic lesions	1	100.0	0.02	0.0–0.1	1	40.0	0.01	0.0–0.03	2	58.3	0.02	0.0–0.04
Other, related to the meninges	1	25.0	0.02	0.0–0.1	17	1.0	0.2	0.1–0.3	37	4.5	0.3	0.2–0.4
Lymphomas & hematopoietic neoplasms	1	100.0	0.03	0.0–0.1	7	100.0	0.1	0.02–0.2	97	99.8	0.8	0.7–1.0
Lymphoma	1	100.0	0.02	0.0–0.1	7	100.0	0.1	0.02–0.2	95	100.0	0.8	0.7–1.0
Other hematopoietic neoplasmsd	.	.	.	.-.	.	.	.	.-.	2	91.7	0.02	0.0–0.04
Germ cell tumors, cysts, & heterotopias	9	72.7	0.2	0.1–0.4	14	68.3	0.2	0.1–0.3	4	14.3	0.03	0.0–0.1
Tumors of the sellar region	10	3.4	0.3	0.1–0.4	143	0.2	1.8	1.5–2.1	495	0.4	4.3	3.9–4.7
Unclassified tumors	26	12.8	0.7	0.4–0.9	113	4.4	1.4	1.2–1.7	889	14.9	7.7	7.2–8.2
Not classified in CBTRUS	3	31.6	0.1	0.0–0.2	6	20.0	0.1	0.01–0.1	15	26.4	0.1	0.1–0.2
Total	213	67.1	5.5	4.8–6.3	758	34.2	9.6	8.9–10.3	4303	34.4	37.4	36.3–38.5

^aRates are age-standardized to the 2011 Canadian standard population.

^bDefined as per the Central Brain Tumor Registry of the United States.¹³

^cNOS = not otherwise specified.

^dEstimates based on fewer than 5 observed cases over the 6-year period are suppressed (denoted with a “-”).

Of 24 unique categories, 9 histology groups were exclusively malignant tumors (29% of reported CNS tumor diagnoses). Eleven categories include tumors with both behaviors (from 5% to <100 malignant) and account for 5% of reported diagnoses. An additional 45% of diagnoses populated 4 categories of primarily nonmalignant tumors (<5% malignant). The remaining 19.5% of CNS cases are unclassified (Table 1).

Survival

A total of 31 165 cases are included in survival estimates. Survival rates varied widely by tumor histology (Tables 3–5) with the lowest 5-year rates for glioblastomas (7.3%; 95% CI: 6.6%–8.2%) and the highest for tumors of the cranial nerves and spinal cord (99.1%; 95% CI: 97.1%–99.7%). Categories with all malignant tumors had a wide range of 5-year survival rates with a low for glioblastomas (above) to a high for pilocytic astrocytomas (96.8%; 95% CI: 93.9%–98.4%). Categories with <5% malignant tumors (cranial and spinal, meningioma, other meninges, sellar region) have 5-year survival rates above 90%. Survival for all CNS tumors was higher among females (70.2%; 95% CI = 69.3%–71.0%) relative to males (60.4%; 95% CI = 59.4%–61.4%) (Table 3).

Table 3.

One- and 5-Year Net Survival Rate (NSR), 95% Confidence Intervals (CI), and Percent Malignant (% M) for Patients With Primary CNS Tumors (All Ages) By CBTRUS Histology From 4 Provinces (British Columbia, Alberta, Manitoba, and Ontario) 2010–2015

	Total						Male				Female
Histology Group (Specific)a	Total	% M	1 Year NSR %	95% CI	5 Year NSR %	95% CI	Total	% M	5 Year NRS %	95% CI	Total	% M	5 Year NSR %	95% CI
Pilocytic astrocytomac	355	100.0	.	96.7–99.5	96.8	93.9–98.4	180	100.0	.	93.4–99.2	175	100.0	95.9	90.5–98.2
Diffuse astrocytoma	495	100.0	70.0	65.7–73.8	48.9	44.1–53.5	290	100.0	48.4	42.1–54.4	205	100.0	49.7	42.2–56.8
Anaplastic astrocytoma	430	100.0	61.5	56.8–65.9	37.8	32.7–42.9	235	100.0	34.9	27.8–42.1	195	100.0	41.4	34.0–48.5
Unique astrocytoma variantsc	95	70.5	79.3	69.4–86.2	65.2	54.2–74.1	60	69.0	61.7	47.2–73.3	.	73.0	.	.-.
Glioblastoma	5450	100.0	42.2	41.0–43.5	7.3	6.6–8.2	3230	100.0	6.8	5.9–7.9	2220	100.0	8.1	6.8–9.4
Oligodendroglioma	400	100.0	92.2	89.1–94.5	79.4	74.3–83.5	235	100.0	79.4	72.5–84.7	160	100.0	79.4	71.4–85.4
Anaplastic oligodendroglioma	335	100.0	80.0	75.3–84.0	50.1	44.0–55.9	195	100.0	51.8	43.6–59.3	140	100.0	47.8	38.5–56.6
Choroid plexus tumorsc	60	11.3	.	.-.	.	83.6–98.8	.	11.5	.	.-.	.	11.1	.	.-.
Oligoastrocytic tumors	405	100.0	82.0	77.8–85.4	53.1	47.8–58.1	230	100.0	55.4	48.3–61.9	175	100.0	50.2	42.1–57.8
Ependymal tumors	550	59.1	96.2	94.0–97.5	87.9	84.1–90.9	300	58.1	88.4	82.7–92.3	250	60.2	87.3	81.7–91.4
Glioma malignant, NOSb	700	100.0	61.6	57.8–65.1	43.5	38.5–46.4	370	100.0	45.3	39.6–50.8	330	100.0	39.2	33.7–44.7
Neuronal & mixed neuronal-glial tumors	455	17.4	95.7	93.3–97.3	85.1	80.9–88.4	265	18.6	87.4	81.9–91.3	190	15.8	81.8	74.6–87.1
Tumors of the pineal regionc	50	67.3	.	78.5–96.2	71.1	55.1–82.2	.	64.3	.	.-.	.	70.8	.	.-.
Embryonal tumors	300	97.7	81.8	76.9–85.7	66.6	60.7–71.8	185	98.4	64.3	56.6–71.0	115	96.6	70.1	60.7–77.8
Other neuroepithelial tumorsc	.	.	.	.-.	.	.-.	.	.	.	.-.	.	.	.	.-.
Cranial nerves & spinal cordd	2445	0.9	99.1	98.4–99.5	99.1	97.1–99.7	1200	0.8	100.0	.-.	1245	1.0	98.5	96.0–99.4
Meningioma	7530	2.1	93.4	92.7–94.0	86.6	85.3–87.7	2255	3.1	83.3	80.7–85.6	5275	1.7	88.0	86.5–89.3
Mesenchymal tumors	165	35.5	93.3	88.0–96.3	79.4	71.5–85.4	80	42.5	72.4	59.6–81.8	85	29.1	85.9	75.1–92.2
Other neoplasms related to the meninges	325	3.7	96.7	93.7–98.3	92.0	86.8–95.2	165	4.3	92.0	82.8–96.4	165	3.1	91.8	84.6–95.7
Lymphoma	605	100.0	54.7	50.7–58.5	38.2	33.9–42.5	340	100.0	40.8	34.9–46.5	265	100.0	35.0	28.8–41.3
Germ cell tumors, cysts, & heterotopiasc	155	63.0	94.9	90.0–97.5	94.1	88.6–97.0	110	71.6	93.9	87.2–97.2	.	42.2	.	.-.
Tumors of the sellar region	3875	0.4	97.8	97.2–98.3	96.3	94.9–97.3	1965	0.6	96.1	93.9–97.5	1905	0.3	96.5	94.6–97.7
Unclassified tumors	5810	12.3	71.2	69.9–72.4	60.3	58.6–62.0	2470	13.4	56.3	53.5–58.9	3340	11.4	63.3	61.0–65.5
Not classified in CBTRUS	140	25.7	90.5	84.0–94.4	82.3	73.8–88.4	75	28.6	84.2	71.1–91.7	65	22.2	80.4	67.4–88.7
Total	31165	35.7	78.2	77.8–78.7	65.5	65.0–66.3	14505	43.8	60.4	59.4–61.4	16660	28.6	70.2	69.3–71.0

^aDefined as per the Central Brain Tumor Registry of the United States.¹³ Survival rates are shown for selected subgroup categories.

^bNOS = not otherwise specified.

^cValues were suppressed due to one or more of the following reasons: The number of cases at risk was less than 50; less than 5 deaths in the category.

^d100.0 indicates the estimated net survival rate is greater than 100%.

Table 5.

One- and 5-Year Net Survival Rate (NSR) and 95% Confidence Intervals (CI) by CBTRUS Histology for Adult Patients (40+) With Primary CNS Tumors From 4 Provinces (British Columbia, Alberta, Manitoba, and Ontario) 2010–2015

	All Adults (40+)					40–59			60+
Histology Group (Specific)a	Total	1 Year NSR %	95% CI	5 Year NSR %	95% CI	Total	5 Year NRS %	95% CI	Total	5 Year NSR %	95% CI
Pilocytic astrocytomac	50	.	84.2–99.3	88.9	70.8–96.1	.	.	.-.	.	.	.-.
Diffuse astrocytoma	305	56.2	50.5–61.5	30.6	25.1–36.3	145	50.1	41.3–58.2	155	11.4	6.1–18.5
Anaplastic astrocytoma	280	51.2	45.4–56.8	23.8	18.1–29.8	140	33.4	24.9–42.2	140	14.2	7.4–23.2
Unique astrocytoma variantsc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Glioblastoma	5145	40.6	39.3–41.9	6.2	5.4–7.0	1650	9.7	8.2–11.4	3495	4.6	3.8–5.6
Oligodendroglioma	235	87.2	82.1–90.9	72.5	65.5–78.4	180	80.9	73.4–86.5	55	44.5	28.4–59.3
Anaplastic olgiodendroglioma	255	76.8	71.0–81.5	43.7	36.8–50.4	160	59.4	50.3–67.3	100	19.2	10.9–29.1
Choroid plexus tumorsc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Oligoastrocytic tumors	265	73.2	67.4–78.2	39.0	32.7–45.2	155	54.5	46.0–62.2	105	15.3	8.2–24.4
Ependymal tumors	315	94.8	91.3–96.9	88.2	82.4–92.2	195	89.7	83.5–93.7	115	85.8	72.6–93.0
Glioma malignant, NOSb	395	45.0	40.2–49.8	25.7	20.8–30.8	140	44.9	35.9–53.5	255	16.1	10.7–22.4
Neuronal and mixed neuronal-glial tumors	140	89.4	82.6–93.6	65.3	54.9–73.9	85	70.7	58.7–79.8	55	56.6	37.4–71.9
Tumors of the pineal regionc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Embryonal tumorsc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Other neuroepithelia tumorsc	.	.	.-.	.	.-.	.	.	.-.	.
Cranial nerves and spinal cord	1890	99.1	98.2–99.5	99.2	95.7–99.8	1055	.	96.1–100.0	835	98.8	87.9–99.9
Meningioma	6985	92.9	92.2–93.6	85.8	84.4–87.0	2735	93.3	92.0–94.3	4255	80.9	78.8–82.8
Mesenchymal tumors	135	93.9	87.8–97.0	81.5	72.5–87.9	80	80.2	68.8–87.7	50	83.8	65.4–92.9
Primary Melanocytic lesionsc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Other, related to the meninges	220	96.1	91.8–98.1	90.7	83.3–94.9	125	91.7	83.1–96.0	95	89.9	73.1–96.4
Lymphoma	560	53.0	48.9–57.0	35.9	31.5–40.4	150	53.2	44.3–61.2	410	29.6	24.6–34.7
Other hematopoetic neoplasmsc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Germ cell tumors, cysts, and heterotopiasc	.	.	.-.	.	.-.	.	.	.-.	.	.	.-.
Tumors of the sellar region	2955	97.2	96.4–97.9	95.4	93.6–96.6	1490	97.4	96.1–98.3	1470	93.3	89.8–95.6
Unclassified tumors	4995	67.1	65.7–68.4	55.0	53.1–57.0	1245	75.1	72.5–77.6	3750	48.4	46.0–50.8
Not classified in CBTRUSc	85	88.0	78.5–93.4	75.0	62.6–83.8	55	83.4	69.9–91.2	.	.	.-.
Total	25375	74.5	74.0–75.1	60.9	60.1–61.6	9920	73.4	72.5–74.4	15455	53.0	51.9–54.0

^aDefined as per the Central Brain Tumor Registry of the United States.¹³ Survival rates are shown for selected subgroup categories.

^bNOS = not otherwise specified.

^cValues were suppressed due to 1 or more of the following reasons: the number of cases at risk was less than 50; less than 5 deaths in the category.

The overall 1-year survival rates for children and AYA were 90.2% (95%CI = 88.4%–91.7%) and 96.0% (95%CI = 95.4%–96.5%), respectively. However, rates declined at 5 years post-diagnosis to 81.5% (95%CI = 79.2%–83.6%) for children and 87.9% (95%CI = 86.9%–88.9%) for AYA (Table 4). A greater number of tumor types are diagnosed in AYA relative to children, as they are diagnosed with tumors that are classified as both pediatric and adult (Table 4). Estimated 5-year survival rates for most adult tumor categories were significantly different (Table 5). Overall 5-year survival in the 40–59-year-old group was higher (73.4%; 95%CI = 72.5%–74.4%) than in the 60+ group (53.0%, 95%CI = 51.9%–54.0%).

Table 4.

One- and 5-Year Net Survival Rate (NSR), 95% Confidence Intervals (CI), and Percent Malignant (% M) by CBTRUS Histology for Children and Adolescents/Young Adults (AYA) With Primary CNS Tumors From 4 Provinces (British Columbia, Alberta, Manitoba, and Ontario) 2010–2015

	All Children (0–14)						All Young Adults (15–39) Years
Histology Group (Specific)a	Total	% M	1 Year NSR %	95% CI	5 Year NSR %	95% CI	Total	% M	1 Year NSR %	95% CI	5 Year NSR %	95% CI
Pilocytic astrocytomac	195	100.0	.	96.0–99.8	.	95.9–99.8	110	100.0	.	93.3–99.9	.	89.6–98.7
Diffuse astrocytomac	50	100.0	.	78.5–95.9	82.5	69.0–90.5	140	100.0	93.6	88.0–96.6	76.9	68.2–83.6
Anaplastic astrocytomac	.	100.0	.	.-.	.	.-.	135	100.0	84.3	76.9–89.5	68.3	59.1–75.8
Unique astrocytoma variantsc	.	42.1	.	.-.	.	.-.	.	62.5	.	.-.	.	.-.
Glioblastomac	.	100.0	.	.-.	.	.-.	265	100.0	77.5	72.0–82.1	28.6	22.7–34.8
Oligodendrogliomac	.	100.0	.	.-.	.	.-.	160	100.0	.	95.1–99.9	89.7	82.1–94.1
Anaplastic olgiodendrogliomac	.	100.0	.	.-.	.	.-.	75	100.0	90.9	81.7–95.6	70.4	57.9–79.9
Choroid plexus tumorsc	.	19.4	.	.-.	.	.-.	.	8.3	.	.-.	.	.-.
Oligoastrocytic tumorsc	.	100.0	.	.-.	.	.-.	140	100.0	.	94.2–99.7	80.9	72.6–87.0
Ependymal tumorsc	85	90.5	.	90.8–99.4	82.4	71.3–89.5	155	58.2	.	94.0–99.4	90.2	83.7–94.2
Glioma malignant, NOSb	155	100.0	78.2	70.8–83.9	59.2	51.0–66.5	150	100.0	90.8	84.9–94.5	70.3	61.8–77.3
Neuronal & mixed neuronal-glial tumorsc	85	3.4	.	91.0–99.5	.	86.7–98.2	230	7.9	.	96.0–99.6	92.8	88.0–95.7
Tumors of the pineal regionc	.	100.0	.	.-.	.	.-.	.	71.4	.	.-.	.	.-.
Embryonal tumors	190	99.0	85.0	79.1–89.3	71.2	64.1–77.2	70	97.1	85.8	75.2–92.1	68.5	55.5–78.5
Other neuroepithelial tumorsc	.	66.7	.	.-.	.	.-.	.	100.0	.	.-.	.	.-.
Cranial nerves and spinal cordc	50	5.9	.	.-.	.	.-.	505	1.6	.	97.7–99.6	98.5	96.5–99.4
Meningiomac	.	12.5	.	.-.	.	.-.	530	2.1	98.9	97.5–99.5	96.7	94.5–98.1
Mesenchymal tumorsc	.	83.3	.	.-.	.	.-.	.	47.6	.	.-.	.	.-.
Other, related to the meningesc	.	25.0	.	.-.	.	.-.	100	1.0	.	92.2–99.5	.	88.3–98.1
Lymphomac	.	100.0	.	.-.	.	.-.	.	100.0	.	.-.	.	.-.
Other hematopoetic neoplasmsc	.	100.0	.	.-.	.	.-.	.	100.0	.	.-.	.	.-.
Germ cell tumors, cysts, and heterotopiasc	50	73.1	.	85.4–99.0	.	83.1–98.2	80	68.3	.	87.5–98.2	.	86.0–97.7
Tumors of the sellar regionc	60	3.4	.	88.5–99.8	.	85.9–99.1	855	0.2	.	97.4–100.0	99.4	98.0–99.8
Unclassified tumors	150	11.3	94.1	88.9–96.9	94.2	88.9–97.0	660	3.3	97.6	96.1–98.6	92.9	90.5–94.8
Not classified in CBTRUSc	.	31.6	.	.-.	.	.-.	.	20.0	.	.-.	.	.-.
Total	1265	67.1	90.2	88.4–91.7	81.5	79.2–83.6	4525	34.1	96.0	95.4–96.5	87.9	86.9–88.9

^aDefined as per the Central Brain Tumor Registry of the United States.¹³ Survival rates are shown for selected subgroup categories.

^bNOS = not otherwise specified, % M = Percent malignant.

^cValues were suppressed due to one or more of the following reasons: the number of cases at risk was less than 50; less than 5 deaths in the category.

Histological Subtypes

The most commonly diagnosed types of CNS tumor are meningioma followed by glioblastoma (Table 1). There was a high proportion of unspecified or unclassified tumors (19.5%) which represent a potentially heterogeneous mix of histological subtypes for which the cancer registries may not have sufficient information or resources to code (Table 1).

Meningioma

Meningioma is the most common CNS tumor (23.9%) and is 2.3 times more common among females than males (Table 1). In a reversal of the general male preponderance of CNS tumors, meningioma male rates (ASIR = 1.6 per 100 000 person-years, 95%CI = 1.5–1.8) for meningiomas are lower than female rates (ASIR = 3.8 per 100 000 person-years, 95%CI = 3.6–4.1). There is variation in the 5-year survival rate across sexes: with male rates (83.3%; 95%CI = 80.7%–85.6%) lower than female rates (88.0%; 95%CI = 86.5%–89.3%) (Table 3). The 5-year survival rate is also lower among those over age 60 (80.9%; 95%CI = 78.8%–82.8%) relative to those aged 40–59 years at diagnosis (93.3%; 95%CI = 92.0%–94.3%) (Table 5).

Glioblastoma

Glioblastoma is the second most common CNS tumor overall (17.3%), and the most commonly diagnosed CNS tumor among males (ASIR = 2.3 per 100 000 person-years; 95%CI = 2.1–2.5) (Table 1). These tumors have poor overall survival rates with 42.2% (95%CI = 41.0%–43.5%) surviving 1-year and 7.3 (95%CI = 6.6%–8.2%) 5 years (Table 3). Rates decline as age increases (Tables 4 and 5). Middle-aged adults have 5-year survival rates (9.7%; 95%CI = 8.2%–11.4%) higher than older adults (4.6%; 95%CI = 3.8%–5.6%). There are too few cases of these tumors in children to report.

Unclassified tumors

A heterogeneous category of “unclassified” CNS tumors included an average of >1000/cases/ year (Table 1). The proportion of unclassified tumors varied by sex (18% in males and 21% in females) and age group (12.2% in children, 14.9% in AYA, and 20.7% in adults) (Table 2). In the AYA group, none of the malignant cases and 21.4% of the nonmalignant cases were unclassified (Figure 1). 1-year survival rates for unclassified tumors (71.2%; 95%CI = 69.9%–72.4%) were lower than overall rates (78.2%; 95%CI = 77.8%–78.7%) a pattern which persisted at 5 years (unclassified 60.3%; 95%CI = 58.6%–62.0% vs overall 65.5%; 95%CI = 65.0%–66.3%). The 5-year survival rates in the unclassified group were lower in males (56.3%; 95%CI = 53.5%–58.9%) than in females (63.3%; 95%CI = 61.0%–65.5%) (Table 3) and declined within adult age groups ages 40–59 years (75.1%; 95%CI = 72.5%–77.6%) and over age 60 years (48.4%; 95%CI = 46.0%–50.8%) (Table 5).

Figure 1.

The percent of unclassified cases by behavior, sex, and age group.

Discussion

We summarize population-level surveillance information on all primary brain tumors in the Canadian context. Detail by tumor behavior is available in concurrent reports elsewhere.¹⁵ The most common histological subtypes were meningioma and glioblastoma. Overall, survival was better among females, relative to males. This may be due in part, to the lower proportion of malignant tumors in females (28.6%) compared to males (43.8%) (Table 3). The variation in incidence and survival rates by age and sex in specific tumor types reflect the heterogeneity and complexity of these tumors, and highlights the need for routine surveillance reporting. The small number of cases per year for many tumor subtypes underscores the need to compile information across regional populations to support tumor-specific research and provide basic information needed to measure future improvements in patient outcomes. Findings are consistent with what was expected, based on evidence generated by the U.S CBTRUS.¹²

The inclusion of nonmalignant brain tumors in cancer registries is a relatively recent occurrence. While argued as a critical need from a research and clinical perspective for many years^5,16 reporting requirements were not embedded into the law until 2003¹⁷ in the United States and 2007³ in Canada. International data on the full spectrum of these tumors is now emerging. This transition in CNS data collection initially underreports tumors, to which Canada is no exception.⁴ We anticipate underreporting will decrease as the specificity in CNS coding becomes the norm and newer strategies for case ascertainment are implemented over time.

The provinces included in the present report contributed to efforts to address underreporting and enhanced case ascertainment processes which are now embedded in cancer registry workflows. Remaining provinces and territories will be included in ongoing surveillance reports using data from the Canadian Cancer Registry. The value of adding the groupings with both malignant and nonmalignant tumor behaviors to the cancer registry becomes apparent in these data; 47.1% of all classified CNS tumor diagnoses have nonmalignant behavior, and mixed behavior is found in 14 of 24 unique histology groups.¹⁸ The high proportion of CNS tumor categories with nonmalignant behaviors underscores the need for high-quality registry data on these specific tumors, so that surveillance and research reflect the true underlying patient population.

We encourage regional support to enhance case reporting to each respective provincial and territorial cancer registry, with consideration for wider access to and adoption of data collection from electronic medical records. Such regional improvements are necessary to ensure that case counts in the Canadian Cancer Registry will become complete and coding will become consistent nationwide over time.⁷ Legislative initiatives recognized that the addition of primarily nonmalignant tumors created an increase in the overall workload of cancer registries counterbalanced by the potential to improve the quality of life for patients with these tumors. For example, as survival rates have improved for pediatric brain tumor patients, a shift to a better understanding of the quality of life among survivors has emerged leading to the development of interventions relating to improving educational outcomes.¹⁹ In adults, a better understanding of the positive and negative effects of treatment on neurocognitive function and health-related quality-of-life measures has emerged.²⁰ Future improvements in quality of life for the nonmalignant group (45% of all CNS patients) require that data are captured appropriately to facilitate clinical research and practice.

The proportion of tumors that are unclassified in this dataset (19.5%) was unexpectedly high. The US reports 5.5% of brain tumors are similarly unclassified.¹² Factors that may influence variations by sex, age, and behavior include: The relatively recent introduction of these tumors into Canadian data systems; variations in the data sources and approaches to identifying and coding these tumors in each provincial registry, and the relatively rapid changes in diagnostic methodologies taking place.^7,21 The proportion of unclassified tumors was highest in Ontario, a registry that has at its disposal non-pathology-proven cases from CIHI discharge and cancer center treatment reports.²² While Ontario has the advantage of identifying incident cases that would otherwise be missed, current data sources do not provide adequate detail for the broad tumor categories used in surveillance reports. Access to other, medically verified sources of data such as imaging and molecular markers tailored to the data structures of each jurisdiction could resolve this issue and allow registries to provide highly clinically relevant information.

Differences in clinical management may also affect pathways to diagnosis, guidelines for laboratory testing, or follow-up care at the provincial level. Given the uncertainty around the composition of this group, the reported incidence rates should be interpreted with caution. This is particularly true for older and nonmalignant CNS cases, as they have higher proportions in the unclassified category (Figure 1). We expect that as case capture for these tumors improves over time, the proportion of unclassified nonmalignant tumors will decrease.

Given changes that have taken place over time with respect to CNS tumor capture in cancer registries and multiple classification changes, we urge caution in the interpretation of overall and histologic-specific trends over time.^21,23 Additionally, a small number of cases that were coded as occurring in the brain did not fit into the CBTRUS classification system. This does speak to gaps that may arise from diagnostic and classification changes and the need for continuous training of cancer registrars. Differences when comparing Canada with other countries may also be explained by differing population age structures as the incidence rates reported are standardized to the Canadian 2011 population and each survival rate estimate is adjusted for the expected survival within the relevant case group age and gender profile. Assessing trends will need to wait while the data accumulate and accuracy improves.

Conclusions

Combining information from these 4 provinces allows closer inspection of data quality and epidemiologic patterns by tumor type than previously feasible in Canada. In spite of data limitations, the incidence patterns by sex and age group for specific histology groups are similar to that reported elsewhere. The potential underreporting of nonmalignant cases and the higher-than-expected proportion of unclassified cases will affect the accuracy of histology-specific information, particularly for tumor groups with mixed behaviors. However, a portion of unclassified tumors are truly unclassified when they are reported to a registry and further assessment of the factors underlying this labeling is warranted.

We emphasize the need to provide support for region-specific approaches for case identification and tumor-specific coding. Focused quality control studies, similar to that conducted in Alberta, Jiang Y., Benign brain tumor surveillance project [unpublished]. Public Health Agency of Canada, Chronic Disease Surveillance Division; 2009. could be used to identify gaps and prioritize promising approaches to improving data processes at the regional level. Case identification may be improved by incorporating natural language processing approaches into radiology and pathology record searches for cases. Algorithms could be developed and modified for each region to maximize the potential information gathering with minimal individual registrar involvement. The coding of tumor type may also be enhanced with focused training and/or ensuring clinical, epidemiological expertise is available when case-specific questions arise. Working together, high-quality population-based data on all primary CNS tumors will increasingly be available to guide research, clinical and policy decisions to improve patient outcomes in Canada.

Conflict of Interest Statement

None declared.

Funding

This work was supported by Brain Canada through Health Canada and Brain Tumor Foundation of Canada.

Author Contributions

Conceptualization: F.D., E.W., Y.Y, Data acquisition, and ethics approvals: F.D., T.S, Data analysis: T.S., E.W., F.Y, Table preparation: F.D., F.Y. Writing and interpretation: F.D., E.W., Y.Y.