Artificial Intelligence (AI) continues to reshape healthcare, and one of its most promising applications is in radiology. Quick and accurate diagnosis is critical, especially for time-sensitive conditions such as pulmonary embolism (PE). Recent research highlights how AI triage software can accelerate radiology report turnaround times, potentially improving patient outcomes—but it also reveals important limitations to consider.
How AI Triage Enhances Radiology Efficiency
A study conducted by the U.S. Food and Drug Administration (FDA) in collaboration with the University of Chicago focused on the use of AI triage software on CT pulmonary angiography (CTPA) scans. The software, BriefCase by Aidoc, was designed to automatically prioritize scans showing urgent findings, alerting radiologists to cases that need immediate attention.
The results were compelling: during standard work hours, turnaround times for radiology reports dropped by 32.2%, from approximately 68.9 minutes to 46.7 minutes. This improvement is particularly impactful in acute conditions like PE, where every minute can affect treatment outcomes. By highlighting critical cases more quickly, AI triage software allows radiologists to focus on high-risk patients without delay.
Limited Impact During Off-Hours
Despite these promising results, the benefits were less significant outside of regular work hours. During nights and weekends, report turnaround times saw only a 6.3% reduction, decreasing from 44.8 minutes to 42 minutes.
This suggests that while AI can help prioritize urgent cases, its effectiveness is influenced by other factors, such as staffing levels, radiologist availability, and overall workflow during off-hours. Essentially, AI alone cannot fully compensate for reduced human resources.
Workflow Factors That Influence AI Effectiveness
To better understand the impact of AI triage, researchers used a computational model called QuCAD to simulate different clinical scenarios. The model revealed several key factors that determine how much time AI can save:
- Examination interarrival time – The frequency of incoming scans affects how quickly critical cases are noticed.
- Number of radiologists – More radiologists can offset some benefits of AI, while fewer staff make AI prioritization more valuable.
- Read time per scan – Shorter reading times amplify AI benefits by allowing faster handling of prioritized cases.
- Disease prevalence – AI is more effective when the proportion of urgent cases is higher.
- AI diagnostic performance – Accuracy and reliability of the software directly impact its utility.
This analysis emphasizes that AI triage is not a one-size-fits-all solution. Its benefits are maximized when paired with optimized clinical workflows and thoughtful operational planning.
Key Considerations for Healthcare Providers
For healthcare organizations considering AI triage software, several factors should be carefully evaluated:
- Integration with Existing Workflows
Ensure AI tools fit seamlessly into the current radiology workflow to avoid delays or disruption. - Staffing Levels and Coverage
Assess whether AI can meaningfully support off-hour operations, or if additional staffing adjustments are needed. - Continuous Monitoring and Evaluation
Regularly review AI performance, tracking both accuracy and impact on turnaround times, to ensure ongoing efficiency gains. - Training and Education
Radiologists and clinical staff should be trained on interpreting AI alerts to fully leverage the software’s prioritization capabilities.
Conclusion
AI triage software represents a promising advancement in radiology, especially for enhancing report turnaround times during busy periods. By prioritizing urgent cases, it helps radiologists focus on patients who need immediate attention, potentially improving outcomes in life-threatening conditions like pulmonary embolism.
However, the technology is not a complete substitute for staffing and workflow optimization. Its effectiveness depends on thoughtful integration, continuous evaluation, and consideration of operational factors, particularly during off-hours.
Healthcare organizations looking to adopt AI triage should approach implementation strategically to fully realize its potential benefits.
Read the full study here: Radiology Business
