Ken Dec, Chief Marketing Officer, mTuitive
Everyone is talking about AI in oncology. Digital pathology. Machine learning models that can detect HER2-low expression, stratify recurrence risk, and match patients to clinical trials faster than any human reviewer ever could.
The technology is real. The outcomes are compelling. And yet most cancer programs are not actually ready for it.
Not because they lack ambition. Not because they have not heard the pitch. But because of something far more fundamental: the data underneath their workflows is broken.
The Foundation Problem Nobody Talks About
When a surgeon completes a procedure, or a pathologist signs out a case, where does that information go? For most cancer centers, the honest answer is: somewhere. Maybe into a narrative clinical note. Maybe into a PDF that lives inside an EHR and will never be touched again. Maybe into a registry, after a manual abstractor spends hours trying to pull discrete variables out of free text that was never designed to be queried.
AI does not run on PDFs. Machine learning does not thrive on narrative documentation. Digital pathology platforms cannot reach their full potential when the clinical and procedural data surrounding each case exists in disconnected silos, formatted differently across departments, systems, and sites.
This is the foundation problem. And it does not get solved by buying an AI tool.
Structured Data Is Not a Technology Project. It Is a Strategy.
The cancer programs that are positioning themselves to win in the next five years are not starting with algorithms. They are starting with data architecture.
What does that mean in practice? It means capturing oncology data at the point of care in structured, standardized formats. Not as an afterthought. As a discipline embedded into every surgical case, every pathology report, every post-operative workflow. It means ensuring that data from your EHR, your lab information system, your PACS, and your pathology platforms can be aggregated, normalized, and made available for analysis without a team of people manually re-entering it.
When data is structured at the source, something remarkable happens. That data becomes interoperable. It can flow bidirectionally into state and national registries. It can populate accreditation reports automatically. It can serve as the training substrate that AI models actually need to perform well in your specific patient population.
Structure is not an administrative burden. It is the infrastructure on which everything else depends.
The Accreditation Window Is Closing
Here is the near-term pressure that makes this urgent: the bodies that govern cancer program standards are raising the bar on data completeness and quality. Synoptic reporting is no longer optional. Evidence-based quality indicators are no longer aspirational. They are expected.
For cancer registrars, the challenge is real. Manual abstraction from clinical documentation is time-consuming, error-prone, and increasingly unsustainable as case volumes grow and reporting requirements expand. Programs that rely on unstructured documentation to meet accreditation standards are already feeling the strain.
The programs that solve this problem are not just checking a compliance box. They are freeing their registrars to do what actually matters: quality assurance, outcomes analysis, and supporting clinical research. That is a different kind of team. And a different kind of cancer program.
Why This Matters for Digital Pathology Specifically
Digital pathology adoption is accelerating. Whole-slide imaging. AI-assisted biomarker scoring. Computational models that can find prognostic signals invisible to the human eye. Comprehensive cancer centers are announcing full digital pathology transitions on two-to-three year timelines.
But here is what the pathology AI conversation consistently underplays: the imaging layer is only as powerful as the data context surrounding it. An AI model interpreting a whole-slide image needs to understand what surgical procedure was performed, what the pre-operative findings were, what the staging context is, and how this case compares to your institution's historical population. That context lives in your structured clinical and operative data. Or it does not live anywhere useful.
The institutions achieving the most with computational pathology are the ones that treated structured data capture as a prerequisite, not an afterthought. They built the connective tissue first.
What 'Ready' Actually Looks Like
A digitally transformed cancer center is not one that has deployed the most sophisticated tools. It is one where data flows without friction, from the OR to the pathology lab to the registry to the quality dashboard to the research team, in a format that every system can understand and act on.
It is one where surgeons and pathologists spend their time on clinical decisions, not documentation gymnastics. Where compliance reporting is near-automatic rather than an annual scramble. Where a new AI-powered workflow can be evaluated and adopted because the underlying data infrastructure is already there to support it.
Getting there requires a deliberate investment in how oncology data is captured and organized, starting with the most foundational clinical touchpoints. It requires technology that bridges legacy systems rather than replacing them wholesale. And it requires treating data quality as a clinical quality issue, because ultimately, it is.
The question for every cancer program leader is not whether AI and digital pathology are coming. They are already here. The question is whether your data foundation will be ready when they arrive.
mTuitive helps cancer programs turn pathology and operative reports into standardized, AI‑ready data that improves quality, compliance, and accreditation readiness across oncology workflows. With more than two decades of leadership in synoptic reporting and deep integrations with CAP, CoC, and leading LIS/EHR platforms, mTuitive enables providers to deliver safer, more consistent care while building the data foundation for future diagnostics and therapies.
