The Problem Worth Solving
Radiologists read thousands of reports. Buried in the language of those reports are findings that weren’t the point of the scan — a small lung nodule here, a pancreatic irregularity there. Clinically significant? Maybe. Easy to miss in a busy workflow? Absolutely.
These incidental findings don’t always trigger automatic follow-up. They can slip through the cracks between departments, between systems, between appointments. For some patients, that gap costs time they don’t have.
Enter Eon
Baptist Health Herbert Wertheim Cancer Institute is now running an AI platform called Eon specifically to close that gap.
Eon reads the body of radiology reports and cross-references data from the electronic medical record (EMR). It’s not generating diagnoses — it’s doing the unglamorous but critical work of noticing things that need noticing and making sure someone acts on them.
Currently deployed for incidental lung and pancreas findings, with plans to expand into breast, thyroid, kidney, and liver.
Step 1: Scan the Reports
Eon ingests radiology report text and EMR data, looking for language that signals an incidental finding worth flagging.
Step 2: Organize by Concern Level
Findings aren’t treated equally. Eon sorts them by urgency — higher-concern findings get flagged through established clinical processes at the Institute, while lower-concern findings enter a monitored follow-up queue managed within the platform.
Step 3: Close the Loop
Patients and their ordering physicians are notified. The physician then decides next steps based on the individual patient’s context. Eon hands off cleanly — it doesn’t try to make the call.
That last part matters. The platform is explicitly designed as a decision support tool, not a decision-maker.
What Baptist Health Is Actually Getting
The value here isn’t dramatic. It’s operational — and that’s exactly the point.
Care teams spend less time manually hunting through reports for findings that might need follow-up. Findings that previously could have been delayed or missed now have a structured path to action. Patients get contacted. Physicians get organized information rather than noise.
As Dr. Leonard Kalman, Acting System Chief Executive of Baptist Health Cancer Care, put it: “early detection, providing coordinated, timely care and helping patients receive the right care at the right time.”
That’s not marketing language. That’s a description of a workflow problem being solved at scale.
Limitations Worth Noting
Eon doesn’t replace radiologist judgment — it works downstream of it. If a finding isn’t documented clearly in the report text, the platform can only work with what’s there.
The current scope is also narrow by design: lung and pancreas findings only, for now. Expansion to other organ systems is planned, but the rollout is deliberately staged. That’s a reasonable approach for a clinical environment, even if it means the full value takes time to materialize.
And like any AI tool embedded in a healthcare workflow, its effectiveness depends heavily on how well it integrates with existing EMR systems and how consistently care teams engage with its outputs.
The Takeaway
Eon isn’t trying to be the AI that detects cancer. It’s trying to be the AI that makes sure someone follows up when there’s a reason to.
That’s a quieter problem to solve. It’s also, arguably, the more important one.
For healthcare systems evaluating AI tools, this case is a useful benchmark: the highest-impact deployments often aren’t the flashiest ones. They’re the ones that patch the gaps in a process that already exists — and do it without asking clinicians to change how they think.
Observe the workflow. Fix the gap. That’s the job.
Comments (0) No comments yet
Want to join this discussion? Login or Register.
No comments yet. Be the first to share your thoughts!