Three-Hospital Integrated Health System (unnamed)

Emergency Department radiologists faced high volumes of head CT exams requiring urgent triage, with intracranial hemorrhage (ICH) cases demanding rapid prioritization. Without AI-assisted triage, critical ICH studies could be delayed in standard review queues, potentially slowing time-to-diagnosis for stroke and hemorrhage patients.

Aidoc's AI-based ICH notification system was deployed across a three-hospital integrated system and integrated directly into PACS workflows. Radiologists used PACS-integrated AI widgets displaying key images and status indicators that flagged AI-positive ICH cases, enabling rapid prioritization of high-risk studies across 14,707 ED head CT examinations over 12 months.

Mean turnaround time (TAT) for new ICH cases was 12.9 minutes, significantly faster than 16.2 minutes for false positive cases and 15.7 minutes for AI-negative cases. Differences in TAT between new true positive and false positive cases were statistically significant (p=0.001 after excluding follow-up exams), confirming that radiologists leveraged AI alerts to prioritize critical cases and improve emergency stroke workflows.

• Radiologists actively use AI triage alerts to reprioritize worklists, resulting in measurably faster review of true-positive hemorrhage cases.
• PACS-integrated AI widgets with key image previews and status indicators are an effective mechanism for surfacing urgent findings without disrupting standard workflow.
• Even with a 39% false positive rate, AI-based ICH notifications still produced statistically significant TAT improvements for true positive cases, suggesting the alert value outweighs alert fatigue in this setting.