1584
clinical study

Mechanical Thrombectomy, Artificial Intelligence and the Activation of a Pulmonary Embolism Response Team

Materials & Methods

Mechanical pulmonary arterial thrombectomies performed for acute PE and CT angiogram (CTA) were reviewed over a four-year period. The algorithm to directly notify PERT members was initiated in early 2022. The cohort was divided into pre-implementation and post-implementation groups. Time of CTA acquisition, time of notification of a clinical provider by the reading diagnostic radiologist, and time of thrombectomy were gathered and used to compare times to thrombectomy and time from imaging to notification of a clinical provider between the two groups. Clinical outcomes including hospital length of stay and ICU length of stay were collected. 

Results

A total of 13 patients including six patients in the pre-implementation phase and seven patients in the post-implementation phase. Implementation of the algorithm was associated with a mean decrease in time from CT angiography to mechanical thrombectomy by 40% (17.1 hours vs. 10.1 hours) and time from imaging to notification of clinical provider by 38% (0.8 hours vs. 0.3 hours). Clinical outcomes showed a 23% (13 days vs. 10 days) reduction in the length of stay in the ICU and 18% (17 days vs. 14 days) in overall inpatient length of stay. 

Conclusions

The use of an AI algorithm suggests a significant reduction in time from initial imaging to mechanical pulmonary arterial thrombectomy by screening CT angiograms and directly notifying a PERT of potential candidates.