ARTERY presented a new academic paper: “A Deep Learning-Based Fully Automated Pipeline for Regurgitant Mitral Valve Anatomy Analysis from 3D Echocardiography“
The new paper introduces a cutting-edge deep learning pipeline for precise segmentation of mitral valve annulus and leaflets in the end-systolic configuration. Through the extraction of crucial anatomical landmarks and features, we enable quantitative assessment of regurgitant mitral valve anatomy using 3DTEE data.
A standout feature is the innovative multi-decoder residual UNet architecture that enables precise multi-structure mitral valve segmentation. This feeds into an automated module specifically designed to analyze critical anatomical features, aiding in the identification of coaptation defects.
Our tool guarantees rapid and precise annotation of mitral valve structures and morphological characterization without user interaction. This empowers clinicians in diagnosing mitral regurgitation and facilitates informed decisions regarding surgical or transcatheter repair.
This work stands as a collaboration with Ospedale San Raffaele and contributes to the ARTERY project’s development.
Discover more here
Credits to: Riccardo Munafò, Simone Saitta, Giacomo Ingallina, Paolo Denti, Francesco Maisano, Eustachio Agricola, Alberto Redaelli and Emiliano Votta.
