NAACCR Board Identifies Priority Areas to Propel NAACCR Into the Future

Priority Areas to Propel NAACCR Into the Future

Central cancer registries should expect to benefit from sweeping, transformative new capabilities in electronic health records (EHR), biomedical informatics, and artificial intelligence (AI). As the cancer data ecosystem advances, central cancer registries will need to be prepared and empowered to leverage these advances for better, faster, more meaningful, and sustainable population-based cancer surveillance.

The NAACCR Board has identified priority areas that can best position registries to advance their missions. Traditional objectives include measuring the burden of cancer incidence and informing population-based cancer prevention and control. Advanced and future objectives may also include accelerating cancer research and informing learning health systems that improve clinical care and survival for cancer patients. The Board recognizes that member interest and capacity to pursue objectives beyond population-based cancer surveillance vary across the United States and Canada, but the priorities identified here are fundamental to progress toward these objectives.

Reduced funding, coupled with new data sources that characterize modern cancer screening, diagnostics, treatment, and outcomes beyond survival, compels registries to pursue more efficient electronic data capture. A modernized dataset has great potential to increase the use of registry data and funding from non-traditional sources. Efficient electronic data capture is a broad and ambitious goal, but three pragmatic, high-priority areas have emerged.

These priorities are envisioned in the sequence in which they are most likely to strengthen registry operations: first, improve the quality and completeness of electronic pathology data; second, expand direct EHR-to-registry reporting; and third, apply AI in ways that improve efficiency while protecting data quality and public trust:

1. Advances in electronic pathology (ePath) reporting
2. Electronic health record reporting
3. Responsible use of artificial intelligence

Structured Electronic Pathology Reporting

Initiatives to increase efficient reporting to central cancer registries must begin with electronic pathology (ePath) reporting. Over 95% of all cancers are initially diagnosed by a pathologist. Fortunately, federal agencies in the U.S. and Canada have made impactful investments in ePath reporting for several decades. However, continuing investments are still needed to achieve true population-based ePath coverage. In any given population, most cancers are diagnosed in a hospital-based pathology lab. Direct ePath reporting from hospital labs offers the greatest opportunity to increase electronic reporting for most central registries. Ongoing investments in staff and infrastructure from federal agencies will be needed to meet this objective.

An ongoing challenge with ePath reporting has been the reliance on narrative-free text descriptions to ascertain cancer site, histology, behavior, grade, and other biomarkers. For resections and many biopsies, pathologists routinely record standard elements as defined in checklists promulgated by the College of American Pathologists (CAP). CAP actively promotes and facilitates electronic synoptic reporting checklist elements using CAP Electronic Cancer Protocols (eCPs). Most hospital electronic health record vendors support CAP eCPs for standardized data entry by the pathologist.

By definition, ePath reports transmitted to registries are electronic, but in the U.S., they almost always lack the discrete coded data fields that pathologists recorded using an electronic checklist. This is an astounding and disappointing phenomenon: a pathologist applies expertise and effort to assign standard coded data elements to a diagnosis, but the cancer registry receives only a narrative summary of that report. AI methods have been developed to make a best guess at interpreting the pathologist’s intent from the narrative summary, but countless hours of manual review by Oncology Data Specialists (ODS) are still required to confirm site, histology, behavior, laterality, grade, and biomarkers. The Board believes that NAACCR can help identify the reasons for this disconnect and potential solutions. Increasing structured eCP reporting to central registries would be an “easy” win for registries moving toward more efficient ePath reporting and processing. Canadian provinces have already shown that a significant proportion of cancer cases may be transmitted via structured eCP.

Increased Electronic Health Record Reporting

As progress toward population-based ePath reporting advances, the next critical source of electronic cancer patient data will be facility electronic medical records. The future of central registry reporting envisioned by the U.S. Centers for Disease Control and Prevention (CDC) and the National Cancer Institute (NCI) involves eCP pathology reporting to identify cancer patients and cases, followed by direct EHR querying to complete registry abstracts. As currently envisioned by the National Institutes of Health, Health Level Seven (HL7) Fast Healthcare Interoperability Resources (FHIR), pronounced “fire,” is an enabling technical interface standard needed to make this feasible in the U.S.

NAACCR’s role in defining workflows and standards to facilitate efficient EHR reporting to registries is the next high-priority goal identified by the NAACCR Board. In considering such initiatives, it is evident that the traditional central registry workflow, which relies primarily on receiving fully abstracted cases from hospital facilities, must be reconsidered. As recommended by the NAACCR Modular Records Task Force, modular reporting of screening, incidence, diagnostic workup, treatment, and outcome data as they become available in EHR systems will make surveillance more nimble and responsive to innovative uses of central registry data. Facilities, vendors, state registries, and federal agencies must work together to more tightly integrate standardized EHR, hospital, central registry, and federal reporting. Another motivating factor is the need to advance learning health systems to improve patient outcomes in cancer. Cancer registries can play a central role in achieving this overdue advance. As the sole authoritative source of comprehensive information for all cancer patients in the population, registries can support a brighter future in which integrated registry and clinical data from NAACCR’s representative population of cancer patients are leveraged to better predict outcomes and inform therapeutic options for new cancer patients.

Advance Effective and Responsible AI Use

The explosive advances we are now witnessing in AI stand to transform population-based cancer surveillance. The opportunities to incorporate AI into data acquisition, record consolidation, processing, and data use are seemingly endless and, simultaneously, mind-boggling. Worry is also warranted: if AI is not developed and adopted carefully and thoughtfully, cancer registry data quality, completeness, accuracy, reliability, fairness, and public trust could suffer greatly. However, registries are already seeing the benefits of AI methods. Modeling Outcomes Using Surveillance Data and Scalable AI for Cancer (MOSSAIC) is a collaboration between the National Cancer Institute and the U.S. Department of Energy (DOE). Deep learning algorithms to automatically classify and code cancer ePath reports are in full production in Surveillance, Epidemiology, and End Results (SEER) registries and have drastically reduced the need for manual ODS review of narrative ePath reports. CDC’s eMaRC Plus and language model pipelines deployed at the British Columbia Cancer Registry are also demonstrating increased registry efficiencies in ePath processing. AI applications in EHR case-finding and data extraction are also well underway. NAACCR should play a key role in identifying not only the opportunities for AI, but also the risks, standards, and guardrails needed for effective and responsible use. The need is urgent and growing, but I am confident that, together, NAACCR is prepared to shepherd the AI transformation of central registry operations, prepare an AI-empowered workforce, and ensure data quality and expanded data use.

So get ready, NAACCR members: our future is bright, and our work in the priority areas identified by the NAACCR Board will help ensure that our contributions to understanding, preventing, and reducing the burden of cancer are remarkable.