Cancer collection efforts in the United States provide clinically relevant data on all primary brain and other CNS tumors.

Cancer surveillance is critical for monitoring the burden of cancer and the progress in cancer control. The accuracy of these data is important for decision makers and others who determine resource allocation for cancer prevention and research. In the United States, cancer registration is conducted according to uniform data standards, which are updated and maintained by the North American Association of Central Cancer Registries. Underlying cancer registration efforts is a firm commitment to ensure that data are accurate, complete, and reflective of current clinical practices. Cancer registries ultimately depend on medical records that are generated for individual patients by clinicians to record newly diagnosed cases. For the cancer registration of brain and other CNS tumors, the Central Brain Tumor Registry of the United States is the self-appointed guardian of these data. In 2017, the Central Brain Tumor Registry of the United States took the initiative to promote the inclusion of molecular markers found in the 2016 WHO Classification of Tumours of the Central Nervous System into information collected by cancer registries. The complexities of executing this latest objective are presented according to the cancer registry standard-setting organizations whose collection practices for CNS tumors are directly affected.


Kruchko et al. Collection of all primary CNS tumors in the US
happened with the collection of data on "benign" brain tumors. Collection practices are guided by surveillance stakeholders, but also rely on input from professional organizations such as the American Joint Commission on Cancer (AJCC) and the National Comprehensive Cancer Network. Just as the role of the clinician is critical to the care of the individual patient, so is the information the clinician records in the medical record critical to ensuring cancer surveillance data are accurate and complete. 1 For the clinicians treating patients afflicted with brain or other CNS tumors, this role has expanded with the diagnosis criteria required by the 2016 WHO Classification of Tumours of the Central Nervous System 2 (2016 WHO Classification) and requires close attention to their role in the recording of molecular testing from multiple sources in their patients' medical records. Underscoring the importance of the clinician role, the California law, Assembly Bill No. 2325 (2015)(2016), requires all laboratories to report standardized electronic cancer pathology and molecular marker data to the cancer registry by January 1, 2019. This law will enable the California cancer registry system to receive more timely and complete data on molecular testing that in many cases are completed by multiple laboratories and can result in missing data in the electronic health reports. Laboratories that have data for California are currently busy trying to implement standardized electronic reporting using the College of American Pathologists (CAP) cancer checklist templates or other acceptable standardized reporting to meet the new legal requirement. For the cancer registration of brain and other central nervous system tumors, the Central Brain Tumor Registry of the United States (CBTRUS) is the self-appointed guardian of these values. In 2017, CBTRUS took the initiative to promote the inclusion of molecular markers found in the 2016 WHO Classification into information collected by cancer registries. The complexities of executing this latest objective are presented according to the cancer registry standard-setting organizations whose collection practices for CNS tumors are directly affected.

Collection of Population-Based Information on Brain and Other CNS Tumors in the United States-Malignant and Nonmalignant Tumors
Approximately 2% of new cancers diagnosed in the United States are primary brain and other CNS tumors. These tumors are complex and include more than 100 different histologies. 3 Prior to 2004, central cancer registries (CCRs) were required to report only malignant brain and other CNS tumors. Case reports for benign brain tumors were provided to CBTRUS on a voluntary basis by some CCRs. CBTRUS combined data for malignant and benign brain and other CNS tumors and published detailed reports on the incidence of all primary brain and other CNS tumors by histology, sex, age, race, and Hispanic ethnicity. It demonstrated that data on benign brain tumors can be reported in the United States. Registry collection of benign brain tumors was strongly supported by the clinical community, especially those involved in the treatment of these complex and life-altering diseases. The 4 original contributing CCRs grew to 19 before an amendment to Public Law 102-515 (the Cancer Registries Amendment Act) of 1992 was passed. Public Law 107-260, the Benign Brain Tumor Cancer Registries Amendment Act, 4 mandated the addition of benign brain tumors to the list of cancers required to be reported to registries supported by the CDC/NPCR starting January 1, 2004. To support this change, NAACCR added benign, International Classification of Diseases for Oncology, third edition 5 (ICD-O-3 behavior code/0), and uncertain (ICD-O-3 behavior code/1) to its collection standards for brain and other CNS tumors. The NCI/ SEER program also voluntarily adopted this practice. This change in US surveillance policy continues to have major ramifications in cancer collection practices and has provided a reference standard for the incidence of primary brain and other CNS tumors worldwide (malignant and benign). 6 CBTRUS began the practice of addressing benign and uncertain brain and other CNS tumors as "nonmalignant, " the term that will be used in referring to these tumors herein. National attention to the inclusion of nonmalignant brain and other CNS tumors in collection practices came in 2011 when the American Cancer Society, CDC, NCI, and NAACCR focused on these tumors in their annual update of cancer statistics. 7 This report provided a comprehensive evaluation of incidence for all primary (malignant and nonmalignant) brain and other CNS tumors among adults and children for the years 2004-2007 as well as analyses of longer-term incidence trends and survival rates. This first report on nonmalignant brain tumors based on national statistics noted that nonmalignant tumors represented two-thirds of all adult and one-third of all childhood (age 0-19 years) brain and other CNS tumors. In addition, there was relative stability of the long-term incidence trends of malignant tumors of the neuroepithelial tissue, although variations were observed for histologic groups within this category of tumors. The report noted the heterogeneity and diversity of brain and other CNS tumors histologically, genetically, and therapeutically and called for innovative methods to study these and other rare cancers. However, some national statistics for cancer may include only invasive cancers of the brain and other nervous system. Table 1 provides a profile of the sources for populationbased data for brain and other CNS tumors from those organizations dedicated to monitoring cancer. In addition to the variability in reporting definitions for CNS tumors, some organizations provide summary statistics and not histology-specific data in their public statistical reports, although these data are available in their databases. Like other rare cancers, CNS tumors are not always included with this specificity in standard cancer reports. Data elements are available, although access to some is restricted to research use. Researchers are encouraged to visit the organizations' websites for data availability: (CDC/NPCR) https:// www.cdc.gov/cancer/npcr; (NCI/SEER) https://seer.cancer. gov/; NAACCR) https://www.naaccr.org/cina-deluxe-forresearchers. Survival data are listed as limited, as not all CCRs are able to conduct sufficient case follow-up to allow survival calculations or the organization listed does not include the entire US population. Even though CBTRUS is the only source for yearly population-based statistical information for all primary brain and other CNS tumors by histology, sex, age, behavior, race, and Hispanic ethnicity in the United States, its incidence data are obtained from 47 NPCR and 5 SEER CCRs and dependent on its sources' decision to release certain data elements, and for this reason "to be decided" (TBD) is listed. This applies to molecular markers as found in the 2016 WHO Classification whose collection began as of January 1, 2018. To ensure the completeness and accuracy of their data collection, these new data elements must undergo quality control checks before surveillance organizations will release them to CBTRUS and to researchers. Eventually, these data will be incorporated into public health reports, and for this reason TBD was placed in Table 1. By integrating the biomarkers intrinsically in histology terms, the 2016 WHO Classification provided a way to ensure their collection in cancer registration as these tumors cannot be diagnosed without their integrated molecular marker. The clinical role is the foundation for accurate and complete collection and affects the timeline for the availability of this important, necessary, and critical information.

Collection of Population-Based Information on Brain Molecular Markers
Unlike the process involved in the inclusion of nonmalignant brain and other CNS tumors in cancer registration practices, the inclusion of biomarkers in the 2016 WHO Classification in these practices did not involve a public law. 4 Inclusion of biomarkers was part of a systematic process to update information collected by registries for specific cancers to include site-specific data items that are considered highly clinically relevant. Inclusion of these critical biomarkers for diagnosis and prognosis is rooted in a scientific process that includes discovery, validation, and clinical implementation. 8 Like the inclusion of nonmalignant primary brain tumors, the process also included a conference 9 to reach consensus on the selection of those biomarkers intrinsic to diagnosis, resulting in an update to the 2007 WHO Classification and the first molecular classification for CNS tumors. 10,11 As in the 2007 WHO Classification, the 2016 WHO Classification included the assignment of ICD-O-3 codes necessary for capturing diagnosis in cancer registration. 2 As WHO Classification is a dynamic reference system as evidenced by classification publications for brain and other CNS tumors in 1979,1993,2000,2007, and 2016, allowance for implementation, therefore, must be factored into achieving routine practices. Under the leadership of NAACCR, implementation guidelines and recommendations 12 were created to coordinate a variety of activities necessary to implement new data collection standards in 2018 to address   Table 2. 14 In concert with incorporating the ICD-O-3 codes, SEER, in collaboration with NPCR, NAACCR, and other interested surveillance stakeholders, revised the Solid Tumor Rules. Specifically, the Multiple Primary/Histology rules that guide the recording of cancer codes and determine how many primary cancers to record per patient were updated (https://seer.cancer.gov/tools/solidtumor/). Nonmalignant CNS Solid Tumor Rules 2018 and Malignant CNS Solid Tumor Rules 2018 are listed under the Multiple Primary/ Histology rules and include equivalent and equal terms, ambiguous terms, single and multiple tumors, laterality, and priority order for these tumors. 15 Biomarkers and tissue histology are the first-order priority and are based on pathology reports as recommended by the CAP cancer protocols and the CAP electronic cancer checklist (eCC). 16,17 Adding to the complexities of implementing the surveillance guidelines was the finalization of the AJCC eighth edition. 18 After reiterating that a Tumor Node Metastases classification was not applicable to brain and other CNS tumors, the CNS expert committee worked to provide direction for the diagnosis and management of these primary brain and other CNS tumors with reference to the 2016 WHO Classification and ICD-O-3 Topography. Chapter 72 in the AJCC eighth edition set forth guidelines for clinical endorsement and provided descriptions of the biomarkers found in the 2016 WHO Classification. These molecular markers included isocitrate dehydrogenase 1/2 (IDH1/2) mutations and 1p and 19q chromosomal deletions in its list of Registry Data Collection Variables. 18 At the time of the chapter's writing, the registry data collection variables had been collected by registries as sitespecific factors (SSFs) for brain and included 8 SSFs but not IDH; among them were WHO Grade (SSF1), methylation of OG-methylguanine-methyltransferase (MGMT, SSF4), Chromosome 1p Loss of Heterozygosity (LOH, SSF5), and Chromosome 19q (LOH, SSF6). These SSFs were collected at varying levels by the registries; SEER collected all whereas NPCR required registries to collect SSF1 but allowed voluntary collection of the others. CBTRUS quantified the status of completeness of these factors in a joint project with SEER in 2016 and showed that the collection overall has been improving over time (Fig. 1). 19 To the authors' knowledge, there has been no other published population-based report documenting the collection status of MGMT, 1p, and 19q in the current literature. The inclusion of IDH on the registry list in AJCC confirmed the importance of collecting this biomarker for the clinical care of gliomas. 18 The collection of data on the pediatric population, whether reported in statistical reports overall as ages 0-14 years, ages 0-19 years, or by specific, smaller age groups, comes with a sensitivity shared both by the clinical and the cancer registry communities. Only 3720 new cases of brain and other CNS tumors are expected to be diagnosed in children ages 0-14 years in 2019. The molecular markers in the 2016 WHO Classification found most often in children are known and most have been incorporated into clinical practice. However, it was not until pathologists revised their histology classification to integrate these biomarkers in histology terms that the cancer registry community had a mechanism to include them in cancer collection. Called WHO blue books, these WHO classification manuals provide the authoritative guide for histologies found in ICD-O-3 sites and directly affect collection rules. Pathology confirmation is the first criterion that cancer registrars look for when recording tumor histology, so cancer registry stakeholders are keenly attuned to changes in the WHO classification blue books. Access to clinically relevant population-based data that can support research efforts and influence improvements in patterns of care may also provide clues as to the causes of these devastating tumors.
In 2016, NAACCR established the Site-Specific Data Items (SSDI) Task Force. Their charge included reviewing, updating, and reformatting the SSFs that were already being collected and to consider the addition of a limited number of new data items as the highest priority. During a presentation given at the 2017 NAACCR Annual Meeting, CBTRUS highlighted the dilemma of capturing certain 2016 WHO histology/biomarker classifications having no unique ICD-O-3 codes to the SSDI Committee. As Table 3 shows, few histology/biomarker combinations have unique ICD-O-3 morphology codes, which is the rationale for collecting all brain and other CNS biomarkers irrespective of the assignment of unique ICD-O-3 codes. There were 3 factors that supported this rationale: (1) the new AJCC chapter 18    In collaboration with CBTRUS, the SSDI Task Force proposed the addition of a new NAACCR data item (NAACCR #3816) for brain molecular markers. 20 This new data item permits the collection of data on specific markers needed to define clinically important histological subtypes  that are not differentiated in updated ICD-O-3 codes. By having one data element-brain molecular markersadded under SSDI, the collection process for these new CNS biomarkers was condensed both for the registrars recording these data and the programmers implementing the software changes. Codes for capturing these brain molecular markers are defined in NAACCR's Standards for Cancer Registries Volume II Data Standards and Data Dictionary (Table 4). 21 In-depth coding instructions are available in the SSDI Manual. 20 Both documents provide the clinical rationale for collecting each brain molecular marker.
For CNS histology/biomarker classifications, biomarker testing is intrinsic to diagnosis and is critically important for accurate and complete cancer registration at the patient level. Cancer registrars rely on clinicians to record the results of molecular testing in patient records. The Health Insurance Portability and Accountability Act (HIPPA), PL 104-191, permits the electronic transfer of health records, including pathology reporting, to cancer registries. Use of the CAP eCC by pathologists helps ensure that biomarker testing results are available to the CCR in each state for consolidation into national databases. In 2016, representatives from CAP, NPCR, To support the transfer of the biomarker information, the message and format standards for electronic transmission of the biomarker reports from the pathology laboratories to the CCRs has been under review by the NAACCR Standards for Cancer Registries, Volume V, Pathology Laboratory Electronic Reporting, version 5. 22 Since 2017, the NAACR Volume V Revision Task Force has been working to incorporate guidance focused on recording the results of biomarker testing coming from sources such as the CAP eCC. One of the challenges has been the lack of a unique national patient identifier attached to the tissue specimen undergoing biomarker testing. Recognizing the importance of biomarker reporting in cancer registration practices, the Volume V Task Force continues to work to provide the necessary methodology to achieve this goal. Integrated histology/biomarker classification is required for diagnosis of certain brain and other CNS tumors, making biomarker testing part of clinical practices. It is fortuitous that SEER recognized the importance of collecting select biomarkers in 2010 and continues to collect these as SSDIs.

Current Status and Availability of Biomarkers in US Cancer Surveillance Systems
The intricacies of improving the collection and reporting of primary brain and other CNS tumors that have been elucidated here culminated in the inclusion of the biomarkers found in the 2016 WHO Classification into surveillance practices. Table 1 provides an overview of the current status of surveillance (collection and reporting) of these tumors and how these data are made available to the public and researchers.
As stated previously, CCRs are the primary source of data used by the surveillance organizations shown in Table 1. NPCR and SEER provide an annual joint online publication and public-use data set of the official cancer statistics, entitled "United States Cancer Statistics (USCS). " 23 Nonmalignant brain and other CNS tumors are included in overall incidence reporting for 100% of the United States in USCS by combining CCR data from NPCR and SEER. Other publications, SEER's Cancer Statistics Review, 24 NAACCR's Cancer in North America, [25][26][27][28] American Cancer Society's Cancer Facts and Figures, and the Annual Report to the Nation on the Status of Cancer, 29 report malignant and nonmalignant brain and other CNS tumors in varying degrees.
CBTRUS publishes the CBTRUS statistical reports as supplements to Neuro-Oncology for all malignant and nonmalignant primary brain and other CNS tumors overall and by histology, sex, age, race, and Hispanic ethnicity (available from: http://www.cbtrus.org/reports/reports. htm). Survival data on 100% of the US population are currently not collected or reported by any of the organizations included in Table 1 and are, therefore, marked as limited. With the 2016 WHO Classification biomarker reporting implemented, effective January 1, 2018, surveillance organizations will review the data completeness and accuracy, and, when their standards are met, may make these data elements available to CBTRUS. Mortality data published in each report are provided by the National Center for Health Statistics and do not include histology or biomarker information. However, in some instances, the histology may be imputed based on the underlying cause of death.

Conclusion
Through the cooperation and collaboration of the US-based surveillance stakeholders, much progress has been made to ensure that primary brain and other CNS tumors reflect the current standards guiding clinical care and, as such, include the reporting of biomarkers found in the 2016 WHO Classification. These molecular markers are critical for accurate diagnosis and prognosis for brain and other CNS tumors. On January 1, 2018, the new rules described herein were initiated, and collection of the 2016 Table 4 Coding Rules for Brain Molecular Markers (SSDI NAACCR Data Item #3816) in Collection Practice WHO Classification biomarkers began. Having the biomarker data will eventually enable CBTRUS to measure their use in patient care and will provide an important contribution to understanding health care standards in the United States for brain and other CNS tumors. Together, the cancer surveillance community achieved another milestone in brain and other CNS tumor surveillance.

Funding
Funding for CBTRUS has been provided by CDC under contract no. 2016-M-9030, the American Brain Tumor Association, The Sontag Foundation, Novocure, AbbVie, the Musella Foundation, National Brain Tumor Society, the Zelda Dorin Tetenbaum Memorial Fund, the NCI under contract no. HHSN261201800176P, the Uncle Kory Foundation, and from private and in-kind donations. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of CDC or the NCI. QTO is supported by a Research Training Grant from the Cancer Prevention and Research Institute of Texas (RP160097T).