14482
clinical study

Impact of Artificial Intelligence Triage on Radiologist Report Turnaround Time: Real-World Time Savings and Insights From Model Predictions

Materials & Methods
This retrospective single-center study analyzed 11,252 adult CT pulmonary angiography (CTPA) examinations for suspected pulmonary embolism (PE) performed between 2018 and 2022. Pre- and post-AI deployment periods were compared using an FDA-cleared triage device (Aidoc BriefCase, K190072). Turnaround time (TAT) — defined as the duration from scan completion to preliminary report completion — was assessed separately for work-hour and off-hour cohorts. Workflow data, including exam interarrival and radiologist read times, were modeled using the QuCAD computational framework to predict time savings and assess the impact of workflow parameters on AI performance.

Results
The pre-AI cohort (n=4,694) and post-AI cohort (n=6,558) showed a PE-positive rate of 13.3% and 16.2%, respectively. During work hours, mean TAT decreased from 68.9 minutes (95% CI: 55.0–82.8) to 46.7 minutes (38.1–55.2), yielding a 22.2-minute reduction (p=0.004). Off-hour time savings were smaller and not statistically significant (2.8 minutes, p=0.345). Model-predicted time savings (29.6 minutes for work hours) closely matched clinical results, confirming that workflow parameters such as case volume and radiologist availability significantly influence AI benefit.

Conclusions
AI-based triage significantly reduced report TAT for PE-positive cases during peak workflow hours, improving reporting efficiency without affecting off-hour performance. The study highlights that AI impact depends strongly on clinical workload and radiologist availability, emphasizing the value of integrating computational workflow modeling to set realistic expectations for AI performance.

Explore the Latest AI Insights, Trends and Research