Cancer Registry AI: From Collection to Intelligence
From Cancer Data Collection to Oncology Data Intelligence
“Behind every dataset is a patient. The goal is to ensure their story contributes to better outcomes for themselves and for others.”
Every cancer registry contains thousands of patient stories: diagnoses, treatments, outcomes, and lessons that can improve the future of cancer care. However, the value of that information depends on whether it is complete, trustworthy, and available when cancer programs need it.
Today, much of the industry’s limited oncology data expertise is consumed by the work required to collect and structure that information. Growing case volumes, increasingly complex requirements, and a shortage of experienced Oncology Data Specialists have created a widening gap between capturing cancer data and using it.
Artificial intelligence gives the industry an opportunity to close that gap, not by removing the experts who make oncology data trustworthy, but by extending their capacity and allowing their expertise to create value far beyond traditional abstraction.
The Value and the Limitation of Oncology Data Today
The scale of the need is undeniable. The American Cancer Society projects approximately 2.1 million new cancer diagnoses in the United States in 2026. The same report shows the age-adjusted cancer death rate has fallen 34% from its 1991 peak, a reminder of what becomes possible when scientific discovery, earlier detection, better treatment, and reliable data work together.
Cancer registries are essential to that progress. Their longitudinal data supports accreditation, quality measurement, cancer surveillance, research, community assessment and service-line planning. Registry professionals transform fragmented clinical documentation into structured information that can be compared across patients, populations, and time.
Traditional registry information is often retrospective and should not be confused with the real-time clinical, genomic, or diagnostic information used by oncologists to make immediate treatment decisions. Modernizing oncology data management will not turn every registry into a bedside decision-support system overnight. It can, however, dramatically shorten the distance between patient care and usable information and make trustworthy data available for more valuable purposes.
Oncology Data Expertise is Scarce. Its Impact Doesn’t Have to Be.
Oncology Data Specialists are among the most highly trained data professionals in healthcare. They understand disease sites, staging, treatment, recurrence, follow-up, data standards and the clinical context behind each abstract. Their judgment is what turns extracted information into trusted oncology data.
Yet too much of that expertise is spent navigating records, locating data elements and completing repetitive portions of abstraction. At the same time, cancer programs face backlogs, reporting pressure, increasing complexity and a limited pipeline of experienced professionals. Asking organizations to solve this solely by hiring more people is neither realistic nor sustainable.
The questions we have been asking ourselves are: How can technology extend the reach of Oncology Data Specialists? How can it free these experts to focus on the work that demands their knowledge and judgment? And how can we help the industry accomplish more with the expertise it already has?
What Can AI Do in Cancer Registry Abstraction Today?
Today, AI can credibly reduce the manual effort in cancer registry abstraction – reviewing records and populating routine data elements. Used responsibly, these systems can:
- Identify and organize relevant information from structured and unstructured clinical documentation.
- Prepopulate portions of the cancer abstract for expert review.
- Flag missing, conflicting or low-confidence information.
- Prioritize complex cases and exceptions requiring deeper expertise.
- Standardize workflows and improve consistency across teams.
- Reduce abstraction backlogs and accelerate the availability of oncology data.
In Harmony’s open trials, our Oncology Data Intelligence solution has produced approximately 2× faster abstraction with greater than 90% accuracy. These results are promising, but they should be interpreted appropriately. Performance will vary based on case mix, documentation quality, workflow design, and the specific data elements being abstracted. Technology still requires expert validation, quality governance, and continuous monitoring.
That is not a weakness of the model. It is the model. AI accelerates the work; Oncology Data Specialists protect its integrity.
The Future is Not Simply Faster Abstraction
Productivity matters. Faster abstraction can reduce backlogs, improve reporting timeliness, expand capacity and lower the cost of maintaining a complete registry, but speed is only the first benefit.
The larger opportunity is what happens when Oncology Data Specialists regain meaningful time. Their value goes beyond manually entering required data elements. Their deeper value comes from understanding cancer, treatment, staging, data quality, and how oncology information should be interpreted and applied.
With additional capacity, oncology data professionals can contribute more extensively to:
- Quality studies, outcomes analysis and cancer program performance improvement.
- Commission on Cancer standards, reporting and accreditation readiness.
- Synoptic pathology and operative-report monitoring and compliance.
- Clinical research, supplemental abstraction and custom investigator datasets.
- Specialized cancer and clinical registries.
- Clinical-trial identification and research-support workflows.
- Physician, service-line and community-specific oncology data needs.
- AI validation, data governance and the ongoing monitoring of model quality.
Not every responsibility will fit every professional or every cancer program. Some opportunities will require additional education, specialized training or different operating models. The industry has an obligation to build those pathways rather than simply telling people that automation will create “higher-value work” and expecting the work to materialize on its own.
From Better Data Operations to Better Outcomes
It is tempting to draw a direct line from faster abstraction to improved patient treatment. The truth is more disciplined and ultimately more powerful.
| Progression | What it enables |
|---|---|
| AI-assisted capture | Reduces repetitive chart review and accelerates abstract preparation. |
| Expert validation | Applies specialist judgment to resolve exceptions, verify accuracy, and protect data quality. |
| Timely data availability | Makes trusted oncology data available sooner for reporting, analysis, and operational use. |
| Data activation | Supports stronger quality initiatives, research, clinical planning, and decisions that contribute to better outcomes. |
Better outcomes are the destination, but timely and trustworthy data is the infrastructure. Cancer programs cannot consistently improve what they cannot see, measure, or understand. When data becomes available sooner and can be used more broadly, organizations are better equipped to identify gaps, evaluate performance, support research and allocate resources intelligently.
Will AI Replace Oncology Data Specialists?
The public conversation about AI often begins with automation, efficiency and position elimination. Companies tell people technology may replace much of what they do and then seem surprised when those same people are not enthusiastic about helping implement it.
Skepticism within the oncology data community is understandable. It should not be dismissed as resistance to change. It is a rational response to an industry that has too often discussed the future of work without meaningfully including the people who perform it.
We cannot credibly promise that every responsibility or workflow will remain unchanged. AI will alter oncology data management; however, the future should not be designed exclusively by software companies, executives or investors. Oncology Data Specialists must help determine what technology automates, what always requires expert judgment, how quality is measured and which new responsibilities the profession should own.
The goal is not to make oncology data specialists less important. It is to allow their expertise to matter in more places.
One Technology, Multiple Operating Models
Cancer programs will not all modernize in the same way. Some organizations need a fully managed solution that combines Harmony’s oncology data professionals, operational oversight, quality assurance, and AI-enabled workflows. Others have strong internal teams and want to license the technology so their own professionals can abstract more efficiently.
Both models should be anchored in the same principles:
- Human validation and accountability for data quality.
- Transparent measurement of accuracy, productivity and exceptions.
- Technology configured around the organization’s workflows and data environment.
- A commitment to workforce development, not simply labor reduction.
- A plan to use newly available capacity for meaningful cancer program priorities.
This is what it means to move from oncology data collection to oncology data intelligence. It is not simply a technology purchase. It is a redesign of how people, workflows and data work together.
The Opportunity Ahead
The future of oncology data management is not about collecting more information for its own sake. It is about ensuring that every patient story captured by a cancer program can contribute to something larger: better quality, more relevant research, stronger community programs, smarter operational decisions and a deeper understanding of cancer care and outcomes.
AI can help the industry get there faster. Oncology Data Specialists will determine whether the information remains worthy of trust. Cancer program leaders must create the conditions for that data and the experts behind it to achieve the most impact.
Better data does not automatically create better outcomes. But without timely, reliable and usable data, the path to better outcomes becomes much harder to see.
Behind every dataset is a patient. Our responsibility is to ensure that their story contributes to better outcomes for themselves and others.
How Harmony Healthcare is Helping
Harmony Healthcare combines experienced oncology data professionals with AI-enabled technology to improve abstraction productivity, reduce backlogs, strengthen quality, and help cancer programs activate their data. Organizations can engage Harmony through an AI-enabled managed service or license the technology for use by their internal teams. Learn more about Harmony Healthcare’s AI solutions or start a conversation with our team.
