Key Informatics Sessions from NAACCR 2025 Annual Conference

The NAACCR 2025 Annual Conference featured a lineup of dynamic, thought-provoking speakers who set the tone for the future of cancer registries. A key theme throughout the event was the expanding role of informatics in cancer registry operations, reflecting a growing emphasis on data integration, interoperability, and real-time analytics to enhance cancer reporting, research, and public health outcomes. Presentations ranged from advances in electronic pathology reporting and the integration of artificial intelligence at the registry level to insights from global leaders in health data science. The conference showcased the breadth and depth of informatics innovation shaping the future of cancer surveillance.

If you were unable to attend the conference, missed specific sessions, or want a refresher, here are the key informatics presentations worth reviewing:

Plenary 1 Advancements in Cancer Informatics for Cancer surveillance

• Enhancing SEER to accelerate cancer research.
  Dr. Betsy Hsu, Branch Chief at the National Cancer Institute, presented the MOSSAIC project, which utilizes modular tools to expand AI capabilities in cancer registry workflows, including pre-processing automation and external data integration.
• AI in action: Modernizing cancer registry systems across Canada.
  Dr. Jonathan Simkin, Director of the British Columbia (BC) Cancer Registry, discussed collaborative initiatives across Canada aimed at leveraging AI tools to enhance the timeliness and accuracy of cancer registry data, promoting best practices and innovation.
• USCDI+ Cancer: Use cases and technical frameworks to advance cancer data interoperability.
  Dr. Eric Durbin, representing Umit Topaloglu from the National Cancer Institute, provided an in-depth overview of USCDI+ Cancer. The presentation covered its applications in cancer registry reporting, clinical trial matching, and quality improvement within models such as the Enhancing Oncology Model (EOM). Key technical frameworks discussed included semantic data mapping with NCI, caDSR, FHIR, and OMOP vocabularies to enhance data interoperability.

Concurrent 2.D Advances in Use of Artificial Intelligence in Cancer Surveillance

• Automating prostate staging using natural language processing.
  Lovedeep Gondara from the Provincial Health Services Authority in BC presented on the use of Natural language Processing techniques to automatically derive TNM staging information from pathology reports, enhancing the efficiency and accuracy of prostate cancer data abstraction.
• Understanding causes of abstention in information extraction classification workflows.
  Sayera Dhaubhadel of Los Alamos National Laboratory explored methodologies for analyzing abstention patterns within the MOSSAIC Natural Language Processing system. The research aims to refine training protocols to improve classification accuracy in cancer registry data extraction workflows.
• Using natural language processing to create a virtual tumor repository for cancer research.
  Lovedeep Gondara of the Provincial Health Services Authority presented on the development of an automated pipeline utilizing Natural Language Processing to extract key data elements from cancer pathology reports, supporting the establishment of a Virtual Tumor Repository for research purposes.

Concurrent 3.D Democratization of Data

• Anonymization of differential privacy: Choosing the right method for privacy preservation.
  Lovedeep Gondara from the Provincial Health Services Authority presented a comparative analysis of anonymization and differential privacy techniques. The study focused on evaluating their practicality and effectiveness in safeguarding sensitive information within tumor pathology reports, with implications for data sharing and research compliance.
• Combining synthetic data and federated learning for pathology reports.
  John Gounley of Oak Ridge National Laboratory presented a hybrid learning approach that combines synthetic data generation with federated learning of language models. This methodology aims to enhance model performance while maintaining data privacy in the processing of cancer pathology reports.
• Leveraging locally hosted LLMs to enhance operational efficiency while preserving privacy and security at cancer registries.
  Isaac Hands of the Kentucky Cancer Registry shared findings from efforts to evaluate the use of locally hosted Large Language Model (LLMs) on standard computing hardware. The initiative explores how LLMs can enhance operational efficiency in cancer registries while maintaining strict privacy and security standards.

Concurrent 4.D Use of Standards for Interoperability

• Diving into FHIR and streamlining the death clearance process utilizing vital statistics FHIR bundles.
  Stephen Slack of GC&E Federal presented on the application of FHIR bundles generated by state vital statistics departments to improve the death clearance workflow. The session highlighted advancements in data linkage, update automation, and DCO file generation using standardized FHIR formatted files.
• Enhancing cancer data collection and utilization through structured pathology reporting: An examination of data submission to the California Cancer Registry (CCR).
  Dr. Richard Moldwin of the College of American Pathologists (CAP) examined the role of Electronic Cancer Protocols (eCP) in supporting standardized, complete, and timely cancer data reporting to the California Cancer Registry (CCR). The presentation also addressed key barriers to widespread eCP adoption and discussed strategies for improving structured pathology data submission.
• Integration of NAACCR data elements into the OMOP CDM for large-scale epidemiological research.
  Dr. Christian Reich of Nemesis Health presented on the integration of NAACCR data elements into the OMOP Common Data Model (CDM), supporting the goals of the global OHDSI community to produce real-world evidence for clinical decision-making. The presentation highlighted how aligning NAACCR with OMOP can advance cancer research, while also addressing challenges such as variable harmonization, mixed semantic types in value sets, and the need for more standardized covariate definitions.

Concurrent 5.E Advancing Pathology Reporting and Use

• Evaluation of eMaRC Lite software to determine cancer reportability in New York State.
  Jovanka Harrison of the New York State Cancer Registry presented an evaluation of the CDC’s eMaRC Lite software, focusing on its effectiveness in determining cancer case reportability. The study compared the tool’s performance with SEER*DMS auto-coding results and manual review by registry staff, highlighting its potential role in improving automation and consistency in cancer case identification.
• Advancing cancer registry auto-coding: Automation and consolidation of pathology data.
  Dr. Jonathan Simkin of the BC Cancer Registry presented an approach to advancing cancer registry automation by consolidating multiple pathology reports – often linked to a single tumor – into unified, tumor-level records. The presentation outlined methods for integrating key data fields such as site, histology, behavior, and laterality, emphasizing the role of auto-coding in streamlining pathology data for registry use and improving accuracy in multiple primary case management.
• The cancer PathCHART database (CPC DB): Cancer registry data informatics in action.
  Kathleen Kenna of the National Cancer Institute introduced Cancer PathCHART (CPC), a relational SQL database that integrates and standardizes site and morphology codes across major cancer data standard setters. CPC provides a single source of truth for SMCT, supporting consistency, expert review, and future implementation of ICD-O-4. It’s the first resource to link WHO 5th Edition content with SMCT products, enabling improved accuracy and coordination in cancer data standards.

All presentations from the NAACCR 2025 Annual Conference are available through the NAACCR Education and Training Portal. To view the sessions, you must have an active MyNAACCR account and be a NAACCR member.

Visit: https://education.naaccr.org to log in or create an account.