A battle of the brains: exploring the accuracy of artificial intelligence in detecting brain bleeds on CT scans.
Author
Tang, Victor | The Alfred, Australia
Scientific poster presentation (R2-SPNR-14) at RSNA 2023, 26 – 30. November 2023 in Chicago, US
Purpose
How do two commercially available AI algorithms for intracranial hemorrhage (ICH) compare in a head-to-head comparison?
Method
1,241 non-contrast head CT studies were ground-truthed by independent dual-radiologist review. The sensitivity and specificity of ICH identification of two AI algorithms (including Annalise Enterprise CTB) were assessed.
Results
AI algorithms demonstrated high diagnostic accuracy with sensitivity of 84.9% | 88.1%, specificity of 91.1% | 96.9%, and no significant differences between AI algorithms.
Key Takeaway
The two AI algorithms demonstrated high diagnostic accuracy in a real-world dataset.
Disclaimer
Harrison.ai Radiology Solutions were previously marketed as Annalise.ai solutions.
