AI Reduces CT Turnaround and Wait Times for Positive Pulmonary Embolus

In individuals experiencing an acute pulmonary embolism (PE), prompt treatment (such as starting anticoagulation) is crucial for improving patient outcomes. A study was carried out to assess the impact of using artificial intelligence (AI)-based radiologist worklist reprioritization on the report turnaround times for CT pulmonary angiography (CTPA) examinations that were positive for acute PE. The results indicated that utilizing AI-driven worklist reprioritization led to shorter report turnaround time and waiting periods for PE-positive CTPA examinations.

The study conducted by researchers at the University of Texas Southwestern Medical Center (UTSW, Dallas, TX, USA) involved patients who had undergone CTPA in two periods: before (October 1, 2018–March 31, 2019) and after (October 1, 2019–March 31, 2020) the implementation of an FDA-approved AI tool. This tool was designed to reprioritize CTPA examinations at the top of radiologists' reading lists when acute PE was identified. By analyzing timestamps from electronic medical records and dictation systems, the researchers determined three durations: the waiting period (from examination completion to report initiation), reading time (from report initiation to availability), and overall report turnaround time (the sum of waiting and reading times). The team then compared the durations for PE-positive reports, using final imaging as a reference, between the two periods.

Image: AI has been shown to cut CT turnaround and wait times for positive pulmonary embolus (Photo courtesy of Freepik)

The research discovered that the AI-powered worklist reorganization tool was linked to considerably reduced report turnaround (47.6 vs. 59.9 minutes) and wait times (21.4 vs. 33.4 minutes) for CTPA reports indicating acute PE. However, there was no significant difference in reading times (26.3 vs. 26.5 minutes) for CTPA reports positive for acute PE when compared to reports prior to AI implementation. From these results, the investigators deduced that the AI tool could potentially facilitate quicker diagnoses by radiologists, ultimately leading to earlier interventions for acute PE cases.

Related Links:
UTSW