Improving Follow-Up of Incidental Pulmonary Nodules in the Emergency Department Using an Artificial Intelligence-Supported Workflow.
Automating identification of lung nodules in emergency radiology reports and contacting patients doubled follow-up rates, preventing missed cancers.
This pre-post study demonstrated that implementing an AI-based NLP system to identify incidental pulmonary nodules in emergency department radiology reports and proactively contact patients significantly improved notification (75% → 88.4%) and follow-up rates (53.5% → 67.9%). While no stage shift was observed in the study period, the approach addresses a critical gap in lung cancer early detection by systematically closing the communication loop for incidentally discovered nodules.
What the study was
- Study design
- Retrospective pre-post cohort
- Population
- 228 pre-intervention + 252 post-intervention ED patients undergoing chest CT with incidental pulmonary nodules at University of Maryland Baltimore Washington Medical Center
- Sample size
- 480
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Respiratory Medicine
Why it surfaced
Demonstrates a practical, near-term deployable AI workflow that significantly improves incidental pulmonary nodule follow-up in ED settings; addresses care gap in early lung cancer detection.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.