CARA Lab's segmentation algorithm OCT-AID published in European Heart Journal - Digital Health

In our latest published article, led by Rick Volleberg, we describe that our lab has developed and validated OCT-AID, a deep learning-based algorithm for detailed, pixel-wise analysis of intracoronary OCT images. Trained on over 2800 fully annotated frames from the PECTUS-obs study, the model accurately identifies key anatomical features and plaque types. On both internal and external test sets, OCT-AID showed high agreement with expert assessments, supporting its potential for fast, standardized, and comprehensive OCT interpretation in clinical practice.

Read the full article here!