1455
clinical study

Use of a Machine Learning Algorithm to Detect Incidental Pulmonary Embolus

Materials & Methods 

An artificial intelligence (AI) tool trained to detect PE was applied to a retrospective cohort of 2,632 consecutive contrast enhanced chest CT exams performed at our large healthcare system between Jan. 1 and Feb. 28, 2018. CTA Chest exams using a PE protocol were excluded from the cohort. Natural language processing (NLP) of the cohort reports was used to identify cases where incidental PE was described by the radiologist. All discrepant cases between the algorithm and the NLP results were reviewed by a board-certified cardiothoracic radiologist with more than five years of experience.   

Results

The prevalence of incidental PE was 1.2% (31 cases), lower than the reported 2.6%. The algorithm was 94% sensitive and 99.6% specific in detecting incidental PE. The positive predictive value was 74% and the negative predictive value was 99.9%. Of the 25 discrepant cases that were positive according to the algorithm,14 (56%) demonstrated incidental PE on secondary review. 

Conclusions

The integration of an AI algorithm can improve the detection of incidental PE, expedite interpretation and identify patients that may need immediate medical attention.

Read Study

Aidoc Ask the Experts Webinar

Popular Posts

Article
What Should Hospital Decision Makers Ask AI Vendors?
Article
Infographic: Cultivating the "Giving Tree" of AI
Article
What Does an AI-Powered PERT Workflow Look Like?
Article
Top 3 Takeaways from RSNA 2023

Stay Connected on AI Trends and Insights

Follow us on social

From Our Blog

View all

© Copyright 2023 Aidoc