Diagnostic Performance of Automated Intracranial Hemorrhage Detection Software and Analysis of Volume and Hounsfield Unit Metrics in False-positive Cases

Key Points

• The research aims to evaluate the diagnostic accuracy of automated software for intracranial hemorrhage compared to physician diagnoses and analyze false-positive cases.
• Analyzed 509 head CT cases suspected for intracranial hemorrhage.
• Classified automated detection results as true positive, false positive, true negative, and false negative.
• Compared volume and CT value metrics of false-positive to true-positive cases.
• The software demonstrated 100% sensitivity and 82.2% specificity.
• False-positive rate was found to be 17.8%, with no false negatives.
• Key regions for false positives included sulci, brain parenchyma, blood vessels, and meninges, showing significant statistical differences.