Automated intravascular ultrasound calcium detection

Abstract

Intravascular ultrasound (IVUS) is useful for calcific plaque detection1 and identifying the need forspecialized coronary lesion preparation.2 However, manual labelling requires expertise and specialized software, is time-consuming since it requires frame-by-frame visual inspection, and is subject to variability. Depending on the pullback length, the number of frames can exceed 2,000, rendering this evaluation hard to achieve online during a coronary procedure. We developed an artificial intelligence (AI)-based algorithm for IVUS calcium detection.