Deep learning tool helps radiologists spot brain aneurysms

When assisted by a new deep learning model, radiologists saw a marked improvement in how accurately they detected brain aneurysms from CT angiography imaging, according to a study published in JAMA Network Open this month.

The HeadXNet model was developed by a group of researchers at Stanford University, who trained the algorithm using brain scans from 662 patients. Using this data, the model examines scans and highlights areas in which an aneurysm may have occurred, allowing radiologists to narrow their focus when interpreting imaging data.

In the study, eight clinicians studied 115 head CTA scans with and without the help of HeadXNet. When the algorithm was used, the radiologists demonstrated statistically significant improvements in sensitivity and accuracy, as well as in their agreement with each other's assessments, correctly identifying an average of six more aneurysms per 100 scans than they did without HeadXNet's help.

"There's been a lot of concern about how machine learning will actually work within the medical field," said Allison Park, a Stanford graduate student and co-lead author of the paper. "This research is an example of how humans stay involved in the diagnostic process, aided by an artificial intelligence tool."

More articles about AI:
AMIA encourages stricter FDA oversight of machine learning algorithms
AI tool helps clinicians, parents distinguish between infants' cries
Smartphone app 'listens' to coughs to diagnose respiratory disorders

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months