Registry AI

Registry AI (Cancer Registry Assistant) is a product from Oncora Medical designed for cancer registrars, oncology data specialists, and registry managers at healthcare institutions. It automates core cancer registry workflows by extracting data from pathology reports, radiology studies, and clinical documents, then translating that data into standardized NAACCR coding format. The tool is intended to support not only compliance and reporting obligations but also quality control, analytics, and clinical research data collection.

The system uses a human-in-the-loop approach, connecting to EHR systems via FHIR resources and proprietary data integrations. It ingests data from multiple source types — including Epic FHIR, MOSAIQ/ARIA radiation oncology systems, pathology and radiology reports, and lab results — normalizes those inputs through a data integration layer, and processes them through an AI engine that applies natural language processing to unstructured text, generates NAACCR codes, and runs validation and quality assurance checks.

Core Features

• Intelligent Case Finding: Automatically identifies reportable cancer cases across multiple data sources, with a reported reduction in manual review time of 85% and a lower false positive rate than conventional methods.
• AI-Enabled Abstraction: Extracts and codes key data elements to NAACCR and other formats using FHIR resources and EHR integrations, with reported abstraction accuracy above 95%.
• Follow-up Module: Monitors for critical updates to patient records and connects with state and other external data sources to keep registry data current.
• Analytics and Reporting: Provides insights into cancer program performance and supports reporting to state registries and other entities.
• Quality Assurance: Built-in validation checks identify missing or conflicting data elements to support completeness and accuracy of submitted cases.

AI-Generated Case Forms for Research

• Disease Site Specificity: Generates tailored data collection forms based on specific disease sites and defined research objectives.
• Comprehensive Data Elements: Captures critical data points with intelligent field validation and conditional logic to support high-quality data collection.
• Standardized Terminology: Automatically integrates standard oncology terminology and coding systems to support data compatibility across research platforms.

Research Data Collection Workflow

1. Define Research Goals: Specify research parameters and disease sites of interest.
2. AI Form Generation: The system builds a data collection instrument tailored to the defined parameters.
3. Research Implementation: Deploy the generated forms within the research workflow to collect standardized data.
4. Analysis-Ready Data: Collected data is structured and prepared for statistical analysis.

Data Integration and Processing

• Connects to Epic FHIR for patient demographics, diagnosis codes, and clinical notes.
• Integrates with MOSAIQ and ARIA for radiation oncology data, treatment plans, and dose information.
• Ingests pathology and radiology reports for histology, grade, imaging findings, and TNM staging.
• Pulls lab results including biomarker tests and other diagnostic data.
• A data integration layer normalizes formats across disparate sources before AI processing.
• The AI processing engine applies NLP to unstructured text, generates NAACCR codes, and performs validation and QA checks.
• Outputs are presented in a unified abstraction view showing patient ID, primary site, histology, stage, treatment, and case status.

Registry AI is deployed in connection with existing EHR and oncology information systems and is designed for use within hospital cancer programs and registry departments. The tool supports NAACCR-standard coding and is built to facilitate reporting to state registries and other external entities.