Cancer Registries and Data Management Solutions

‍

Cancer registries play a foundational role in guiding public health programs and policy decisions that shape cancer care and control efforts. 

Registries collect patient data on cancer diagnoses, treatments, and outcomes to create a comprehensive database of how cancer affects populations and how care is delivered across the country. This helps clinicians make informed decisions about cancer prevention, research priorities, and treatment planning. 

Artificial intelligence (AI) and large language models (LLMs) are offering solutions to help manage these large datasets. They support faster, more accurate data collection and advanced analytics that can extract deeper insights from registry data.

‍

Key Takeaways

‍

• Cancer registries are essential databases for cancer diagnosis, treatment, and outcome data nationwide.
• Registry data informs clinical care, outreach, and research, from targeted screening to investigating cancer trends and disparities.
• AI‑enabled platforms like Brim improve data quality and efficiency in registry workflows, allowing registrars to focus on expert review rather than repetitive abstraction.

‍

What is a Cancer Registry?

‍

Cancer registries support public health, research, and clinical care by providing a database of patient notes, treatment plans, and outcomes. Currently, U.S. cancer registries have data on more than 39 million cancer cases since 2001.

Registries are built from reports submitted by medical facilities, hospitals, and pathology labs. These reports are created by clinicians in close collaboration with Cancer Registrars—highly trained data management specialists who collect, organize, and consolidate clinician notes and relevant patient medical records. Registrars ensure consistency and identify any data gaps in each report before sending the information to the state, and ultimately, the central cancer registry. 

‍

Types of Cancer Registry

‍

There are three types of cancer registries. 

1. Hospital-Based Cancer Registries (HBCR)

‍

Hospital‑based registries are facility‑dependent, meaning they are built around the data generated within a specific hospital or group of hospitals.

This type of registry can help inform: 

• Patient care standards and information
• Administrative decision-making
• Education
• Clinical research

To meet specific data needs, hospital‑based registries fall into two subcategories:

1. Single‑hospital registry: Data is collected from one facility only. These registries offer highly specific, locally tailored insights that reflect the unique patient population, practices, and priorities of that institution.
2. Collective hospital registry: Data is aggregated from multiple hospitals. This broader dataset supports benchmarking across institutions, enables comparisons of performance and outcomes, and helps drive system‑wide quality improvement.

Some examples of hospital-based cancer registries include:

• Commission on Cancer (CoC)- accredited hospitals maintain cancer registries with clinical data from more than 1,500 facilities that meet accreditation standards.
• The University of Texas MD Anderson Cancer Institute, a nationally recognized specialty cancer hospital, maintains a collective registry of all of its facilities.
• The Smilow Cancer Hospital at Yale-New Haven maintains a tumor registry for patients treated within its system.

‍

2. Population-Based Cancer Registries (PBCR)

‍

Population‑based registries collect data from an entire population, rather than from a single facility or system. The population can be defined by geography–such as a state or region–or by demographic characteristics such as sex, race, ethnicity, or age. 

This type of data can help inform: 

• Screening and prevention methods
• Rates and trends 
• Treatment plans and outcomes
• Research

Some examples of population-based cancer registries include:

• California Cancer Registry is one of the largest state-wide cancer registries.
• The Gynecologic Oncology Group (GOG) operates a cancer registry to promote research and improve care for cancers that affect women, like ovarian, breast, uterine, and vaginal cancers.
• The National Childhood Cancer Registry is a database of cancer records for patients under age 20 to improve understanding and outcomes for childhood and adolescent cancers.

‍

3. Special Case Cancer Registries

‍

Special case cancer registries collect focused data on a single cancer type or a group of related cancers. This data captures more detailed clinical, genetic, and epidemiologic information than broader cancer registries.

Examples of special case registries in practice include: 

• Roswell Park Familial Ovarian Cancer Registry: Collects data from families with two or more relatives affected by ovarian cancer, supporting work on hereditary risk and family‑based patterns.
• Colon Cancer Family Registry: Gathers clinical, genetic, and family history information from individuals and families with colorectal cancer.
• Lung Cancer Registry: Collected detailed clinical and outcome data for lung cancer cases

‍

Benefits of Cancer Registries

‍

Cancer registries provide data that directly support patient care, guide public health strategies, and inform clinical research. Examples of how cancer registries are supporting real patient care and clinical research include:

• Kentucky cancer registry data show the state had higher-than-average childhood cancer rates, which the state is now investigating through a dedicated research fund
• A Mississippi organization uses cancer registry data to target breast screening outreach and mammogram services to African American women, who are disproportionately affected
• In Massachusetts, data on cervical cancer rates helped improve the screening process to include women over 65

‍

What Data is Included in a Cancer Registry?

‍

Cancer registry data at the population level is a compilation of all the reports submitted by each state. Patient confidentiality is protected across all cancer registry data.

Each report includes:

• Patient demographics: Including age, sex, race, ethnicity, and location
• Cancer characteristics: Such as tumor type, diagnosis site, and stage of disease
• Treatments and outcome: Documenting the treatment course, like surgery or chemotherapy, and the patient’s survival status 

‍

How AI can Support Cancer Registries

‍

Cancer registrars are expert data professionals, but too much of their time gets consumed by the most mechanical parts of the job: hunting through hundreds of pages of unstructured clinical notes to surface the variables that matter. AI and large language models (LLMs) can take on that burden, extracting key clinical details and presenting them to registrars for validation, so their expertise goes toward judgment calls rather than manual lookup.

Not all AI abstraction tools are built the same way, though. The critical difference lies in how a platform handles traceability and trust. Registrars can't simply accept an AI-generated output; they need to verify it, audit it, and stand behind it. That requires a system that shows its work.

Brim Analytics was purpose-built for this workflow. Rather than generating opaque outputs, Brim produces structured, traceable first-pass abstractions grounded in the source record, so registrars can see exactly where each data point came from and efficiently validate or refine it. Brim identifies essential registry variables across diagnosis details, tumor characteristics, treatments, and timelines, and has supported over 100 abstraction projects processing more than 100 billion tokens of clinical documentation. The result is a workflow where AI handles the extraction and registrars do what they do best: expert review, precise coding, and nuanced clinical decision-making.

For cancer registries facing growing case volumes and staffing pressures, this kind of human-in-the-loop approach isn't just more efficient; it's also more defensible.

‍

Bottom Line

‍

Cancer registry abstraction is detailed, high-stakes work. Brim Analytics is built by researchers who understand that, and designed to make your registrars faster without sacrificing the accuracy your data depends on.

If you're curious what that looks like in practice for your registry, we'd love to show you.

‍

Cancer Registry FAQ

‍

Why are cancer registries important for healthcare organizations?

‍

Cancer registries give healthcare organizations a structured, longitudinal view of their cancer cases that EHR data alone can't provide. Clinicians and quality teams use registry data to benchmark against national standards and track outcomes over time. Administrators and researchers rely on it to understand population-level patterns, identify care disparities, and support IRB-approved research. And at the public health level, aggregated registry data shapes screening guidelines, funding priorities, and cancer control policy.

The challenge is that maintaining a high-quality registry is resource-intensive. That's why tools that streamline abstraction without compromising data integrity matter.

‍

‍What is the difference between a hospital-based and a population-based cancer registry?

‍

‍Hospital-based registries track cancer cases within a specific facility or health system, supporting local quality improvement, accreditation, and patient care decisions. Population-based registries collect data across an entire geography or demographic group, making them better suited for understanding cancer trends, disparities, and public health outcomes at scale. Many registrars work within hospital-based systems that report up to state and national population-based registries like SEER.

‍

How long does cancer registry abstraction typically take?

‍

‍Manual abstraction of a single case can take anywhere from 30 minutes to several hours depending on case complexity, documentation quality, and the number of variables required by the registry. AI-assisted abstraction can significantly reduce this time by surfacing relevant variables from unstructured notes as a structured first-pass draft for registrar review.

‍

What variables are typically required for cancer registry reporting?

‍

‍Core variables include patient demographics, primary cancer site, histology, stage at diagnosis, first course of treatment, and follow-up survival status. Specific registries such as CoC-accredited hospital registries or specialty registries focused on oncology subspecialties may require additional variables beyond the standard set.

‍

How do AI tools maintain data quality in registry abstraction?

‍

‍The best AI abstraction tools are designed to support registrars, not replace them. Platforms like Brim generate traceable, source-grounded abstractions that registrars can audit and validate, preserving the accuracy and defensibility that registry reporting requires.