12166
clinical study

Perfecting PERT: AI-Enhanced Clinical Decision-Making in Acute Pulmonary Embolism Interventions at a Large Academic Center

Materials & Methods

This study explores the impact of artificial intelligence on managing acute pulmonary embolism (PE) using an AI-driven workflow. The study employed two AI systems: an AI-driven worklist triage system (Aidoc-PE-only) for alerting on suspected PE in computed tomography pulmonary angiograms (CTPAs), and an AI-driven alert system (Aidoc-PERT) that notified healthcare providers of severe PE cases. The study was divided into three phases: pre-Aidoc (Jan. 2019 to June 2019), Aidoc-PE-only (Jan. 2020 to June 2020), and Aidoc-PERT (Jan. 2023 to June 2023).

Results

Of 6,111 patients who underwent CTPAs during the study period, 71 had PERT interventions. The intervention rate increased from 0.84% during the pre-Aidoc phase to 1.46% during the Aidoc-PERT phase. In the Aidoc-PERT phase, 84% of interventions followed CTPAs and teams were alerted in 89% of cases. However, 16% of interventions were for PE identified on non-CTPA exams, emphasizing the need for AI to address incidental PE as well.

Conclusions

Integrating AI into acute PE management can improve the quality and timeliness of care. The use of AI-driven systems can enhance the triage of severe PE cases, leading to more timely interventions. However, further research is necessary to evaluate the clinical outcomes, such as time to intervention and patient morbidity/mortality, associated with AI-assisted PE management.

Explore the Latest AI Insights, Trends and Research