Incidence Data Sources

Background

Incidence data are from the registries participating in the Centers for Disease Control and Prevention's (CDC's) National Program of Cancer Registries (NPCR) and the National Cancer Institute's (NCI's) Surveillance, Epidemiology, and End Results (SEER) Program. NPCR funds and receives data from 50 cancer registries: 46 states, the District of Columbia, Puerto Rico, the Pacific Island Jurisdictions, and the U.S. Virgin Islands. SEER collects and publishes data from population-based cancer registries in 22 U.S. geographic areas. Data from central cancer registries that are supported by both NPCR and SEER are presented as reported to CDC in 2023.

How incidence data are collected

The primary source of data on cancer incidence is medical records. Staff at health care facilities abstract data from patients' medical records, enter it into the facility's own cancer registry (if it has one), and then send the data to the regional or state registry. Other data sources include physicians' offices, radiation facilities, freestanding surgical centers, and pathology laboratories. Both NPCR and SEER registries collect data using uniform data items and codes as documented by the North American Association of Central Cancer Registries (NAACCR). This standardization ensures that data items collected by the two federal programs are comparable.12 Information on primary site and histology was coded according to the International Classification of Diseases for Oncology, Third Edition (ICD-O-3)3 and categorized according to the revised SEER recodes dated January 27, 2003, which define standard groupings of primary cancer sites. Beginning with 2010 diagnoses, cases are coded based on ICD-O-3 updated for hematopoietic codes based on WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues (2008).

Reportable cases

Cancer is a mandatory, reportable disease for public health surveillance. NPCR and SEER cancer registries consider all incident cases with a behavior code of 2 (in situ, noninvasive) or 3 (invasive, primary site only) in the ICD-O-3 with the exception of in situ cancer of the cervix as reportable. Basal and squamous cell carcinomas of the skin are also excluded, with the exception of those on the skin of the genital organs.3 Several cancers are coded as malignant in ICD-O-3 (beginning with 2001 diagnoses) that were not coded as malignant in ICD-O-23 and are noted as follows:

• Myelodysplastic syndrome (MDS) including refractory anemias (histology codes 9980, 9982–9984, 9989) are included in the "Miscellaneous" and "All Sites" categories.
• Chronic myeloproliferative disease (CMPD) including polycythemia vera and thrombocythemias (histology codes 9950, 9960–9962) are included in the "Miscellaneous" and "All Sites" categories.
• Papillary ependymomas (9393) and papillary meningiomas (9538)—cancers that occur in the central nervous system—are included in the "Brain and Central Nervous System" and "All Sites" categories.
• Some endometrial tumors (8931) are reported in the "Corpus and Uterus, NOS" and "All Sites" categories.

For comparisons with ICD-O-2 for cancers diagnosed prior to 2001, exclude all of the histology codes described above and listed as follows: 8931, 9393, 9538, 9950, 9960–9962, 9980, 9982–9984, 9989, 9990, 9991, 9992.3

Additional changes in ICD-O-3 apply to ovarian cancer: low malignant potential tumors (8442, 8451, 8462, 8472, 8473) of the ovary are no longer coded as malignant. Therefore, these cancers are not accounted for in the calculations of the incidence rate for ovarian cancer included in tables and figures. A footnote is provided as a reminder of this exclusion.

Pilocytic astrocytomas (9421) are also not coded as malignant in ICD-O-3, but these cancers are included in this report.

In situ bladder and breast cancers

In situ bladder cancers were recoded to invasive bladder cancers because the information needed to distinguish between in situ and invasive bladder cancers is not always available or reliable. Counts and rates for in situ breast cancer cases among women are presented; these are reported separately and are not included in counts or rates for the "All Types of Cancer" category.

Unknown sex, age, or race

Non-reportable cancers and cancers in patients of unknown sex or age were omitted from all calculations, but cancers in patients of unknown race were included in the "All Races" category.

Childhood cancer

Incidence data on childhood cancer are published in two formats:

• The first is according to the SEER modification of the third edition of the International Classification of Childhood Cancer. The ICCC-3 is based on ICD-O-3/IARC 2017 classification of Tumors of Haematopoietic and Lymphoid tissues.4 The ICCC presents childhood cancers in 12 groups classified primarily by morphology. The SEER modification, which affects the classification of nervous system and bone tumors, was chosen for compatibility with other published data on rates of childhood cancer in the United States.
• The second format is according to the SEER site recode, which is based primarily on cancer site; the incidence data are presented in this format to make them comparable with published mortality data. This format allows the incidence data for childhood cancers to be categorized in the same groups as adult cancers. Although these groupings are not as appropriate for children as they are for adults, they are necessary to allow comparisons between childhood incidence and childhood mortality.

Nonmalignant brain and CNS tumors

Incidence data on nonmalignant primary brain and central nervous system (CNS) tumors are available in the U.S. Cancer Statistics Data Visualizations tool. Cancer registries began collecting information on nonmalignant brain and CNS tumors beginning with 2004 diagnoses. Data collection of these tumors is in accordance with Public Law 107-260, the Benign Brain Tumor Cancer Registries Amendment Act, which mandates that NPCR registries collect data on all brain and CNS tumors with a behavior code of 0 (benign) and those with a behavior code of 1 (borderline), in addition to in situ and malignant.5 SEER registries voluntarily agreed to incorporate registration of these tumors in their standard practices.5

Effects of events on cancer incidence data

Effect of Hurricanes Katrina and Rita on presenting cancer incidence data

The population of many counties along the Gulf Coast of Louisiana, Alabama, Mississippi, and Texas were displaced in the fall of 2005 by Hurricanes Katrina and Rita, resulting in incomplete case ascertainment for the latter half of the year. For these states, state- and county-level incidence rates were calculated based upon the data submitted to CDC.

Effect of Hurricane Maria on presenting Puerto Rico's cancer incidence data

Puerto Rico's 2017 incidence counts and corresponding rates are based on the first 6 months of the reported data coupled with half of the population estimate (January to June 2017). Cases with an unknown month of diagnosis were also included. To account for the population shift that occurred due to Hurricane Maria in September 2017, data from July to December 2017 are excluded. This population shift may have resulted in incomplete case ascertainment for the latter half of the year.

The population denominators were adjusted by dividing the US Bureau of the Census's July 1, 2017 (vintage 2021) Puerto Rico population estimate in half.

Effect of COVID-19 on cancer incidence data for diagnosis year 2020

In March 2020, the World Health Organization declared COVID-19 a pandemic. Soon after, stay-at-home orders, business and school shutdowns, and travel advisories were implemented in the United States to prevent the spread of COVID-19. Additionally, some health care systems reduced access to routine care. These measures, along with concerns about getting COVID-19, interrupted cancer screening, diagnosis, and care, as people postponed or deferred health care visits, particularly from March to May 2020.

The 2023 data submission includes new cancer cases diagnosed in 2020 and 2021, the first and second years of the COVID-19 pandemic. The missed cancer diagnoses resulting from disruptions in health services caused by the pandemic may have contributed to an observed decline in incidence for most cancer sites in 2020.67 The numbers of new cases diagnosed in 2021 are still a little lower than expected for some cancer types but have returned to pre-pandemic counts for other cancer types. Caution must be taken when examining trends to avoid incorrect interpretations of the effect of cancer prevention and early detection efforts. Observed downward trends may be due largely to the lower observed incidence in 2020.6 CDC and NCI include the 2020 incidence rates in statistical reports and graphics, but do not include them in joinpoint models. The 2021 incidence data will be included in statistical reports and joinpoint models (not included in Data Visualizations tool). JoinPoint software allows researchers to exclude incidence data for 2020, 2021, or both years from trend analyses. Exclude 2020 data for incidence trend analyses, but 2021 data can be included in reports and trend analyses. Learn more about the impact of COVID-19 on cancer incidence trends.