Wednesday, December 3 | 3:30 p.m.-3:40 p.m. | W7-SSGI13-4 | Room E451B
In this talk, researchers will update attendees on their efforts to apply AI to summarize liver contrast-enhanced ultrasound (CEUS) exams and characterize lesions.
Presenter Derya Gol Gungor, PhD, of Siemens Healthineers and colleagues have developed an AI approach that combines arterial phase (AP) washin video frames with portal vein phase (PVP) washout images. These AP washin and PVP washout patterns are then presented in a study summarization tool.
After training, the deep-learning model had 90% accuracy for identifying washout on the test set of 50 lesions, demonstrating its reliability for detecting washout, which is now recognized as the most reliable marker of malignancy, according to the researchers.
“AI-enhanced CEUS imaging transforms liver lesion evaluation, improving diagnostic accuracy and streamlining workflows,” the authors wrote. “This system equips clinicians with advanced tools for better diagnosis and management of liver diseases, ultimately enhancing patient outcomes.”
How did they achieve these results? You’ll need to attend this Wednesday afternoon presentation to find out.
