Healthcare / Health Plans · Radiology & Imaging Operationsprovider
Incidental Finding Follow-Up
Readiness: Now — Deployable with established commercial tools today · Near-term — Proven but early — expect one to three years to mainstream · Emerging — Demonstrated, not yet production-mainstream
Readiness reflects an editorial assessment against a published rubric as of August 2026 — an observation about current tool maturity and adoption, not a prediction about specific products or timelines.
Trajectories describe the observable direction of human effort — not a prediction about specific roles, headcount, or individual careers.
What You Do Today
You dictate the finding you were asked about, and then you dictate the one nobody asked about — a pulmonary nodule on a CT ordered for chest pain, an adrenal lesion on a scan for kidney stones, a thyroid nodule caught at the edge of a cervical spine study. You put the recommendation in the report: repeat imaging at the interval the guidance supports, sometimes with a qualifier about the patient's risk. Then you sign, and the study leaves your hands. Whether that scan ever happens depends on whether the ordering clinician read past the impression, whether the patient came back, and whether anyone owned it after the ED visit that generated the study ended. Nothing on your worklist tells you it didn't happen. The next time anyone hears about it may be years later, when the nodule is no longer incidental.
AI Technologies
Roles Involved
How It Works
NLP reads the narrative of every signed report — not only the impression — and pulls out the sentences that recommend something: a modality, an interval, a condition. Each extracted recommendation becomes a tracked item with an owner and a due date rather than a sentence in a document nobody reopens. The tracking layer then watches orders and scheduling for the study that would satisfy it, and when the due date passes with nothing scheduled, the item surfaces on a worklist a person works — commonly a nurse navigator or follow-up coordinator — who contacts the ordering clinician or the patient. Guideline matching runs earlier, at dictation: as the radiologist describes the nodule or lesion, the reporting template offers the recommendation the applicable guidance supports, so what ends up tracked is specific enough to track.
What Changes
The recommendation stops living only in prose. A finding that used to depend on the ordering clinician reading past the impression becomes a tracked item with a date and an owner. The department gains a view of its own backlog — how many recommendations are open, how many are overdue, which referral sources lose them — which most imaging services do not have today. Patients who fall out of the system because the encounter that generated the study ended get caught by the date rather than by chance.
What Stays the Same
Whether to recommend follow-up at all is a radiologist's judgment, and it is not a neutral one: every recommendation commits a patient to another scan, another wait, and sometimes a biopsy of something that was never going to harm them. Guidance is written for populations, and the patient in front of you may not fit it — a lesion in someone with a known primary is a different question from the same lesion in an otherwise healthy adult. The extraction layer is reading language, and radiologists hedge on purpose: 'could be considered', 'if clinically warranted', 'correlate clinically'. Turning a hedge into a hard due date manufactures an obligation the radiologist did not intend, and someone has to govern that. Whether the follow-up is right for this patient stays with the ordering clinician, who knows things the report does not. The program also needs an owner accountable for closing items rather than a dashboard that counts them — a tracked recommendation nobody chases is the same lost finding with better reporting. And there is a standing conflict to manage, because these programs generate downstream imaging: the clinical indication has to be what drives the recommendation.
Evidence & Sources
- •ACR Incidental Findings Committee white papers on managing incidental findings
- •Fleischner Society recommendations for incidentally detected pulmonary nodules on CT
- •ACR Practice Parameter for Communication of Diagnostic Imaging Findings
- •AHRQ work on diagnostic safety and follow-up of diagnostic test results
Sources listed are directional references, not formal citations. Verify against primary sources before using in business cases or presentations.
Last reviewed: August 2026
What To Do Next
This section won't tell you what your numbers should be. It will show you how to find them yourself. Every instruction below produces a real, verifiable result in your organization. No benchmarks, no projections — just the steps to build your own evidence.
Establish Your Baseline
Know where you are before you move
Before adopting AI tools for incidental finding follow-up, document your current state in utilization management.
Without a baseline, you can't tell whether AI actually improved incidental finding follow-up or just changed who does it.
Define Your Measures
What to track and how to calculate it
patient outcomes
How to calculate
Measure patient outcomes for incidental finding follow-up before and after AI adoption. Pull from your imaging systems.
Why it matters
This is the most direct indicator of whether AI is adding value to utilization management.
clinical documentation quality
How to calculate
Track clinical documentation quality using the same methodology you use today. Don't change how you measure just because you changed how you work.
Why it matters
Speed without quality is just faster mistakes. Measure both together.
Start These Conversations
Who to talk to and what to ask
CMO or VP Clinical Operations
“What's our plan for AI in utilization management? Are we piloting, planning, or waiting?”
This tells you whether to experiment quietly or push for formal investment in incidental finding follow-up.
your imaging systems administrator or vendor
“What AI capabilities exist in our current EHR system that we're not using? Most platforms are adding AI features faster than teams adopt them.”
The cheapest AI adoption is the features already included in your existing license.
a practitioner in utilization management at another organization
“Have you deployed AI for incidental finding follow-up? What worked, what didn't, and what would you do differently?”
Peer experience is more useful than vendor demos. Find someone who has actually done this.
Check Your Prerequisites
Confirm readiness before you invest
Check items as you confirm them.