BeatMapSeq: A Framework for In‐Line Beat‐Based Cardiac Parametric Mapping Sequence Construction and Map Building

ABSTRACT This study aims to propose a framework, named BeatMapSeq, designed to simplify the sequence construction and map building for cardiac parametric mapping in cardiovascular magnetic resonance. BeatMapSeq comprises two components: the BeatMap Framework and DeepFittingNet. BeatMap Framework is a general framework for cardiac beat–based parametric mapping sequences, enabling interactive construction of various cardiac parametric mapping sequences directly on the scanner without the need for programming. DeepFittingNet is a neural network designed for automated in‐line map generation for various parametric mapping sequences. BeatMapSeq was validated on Philips 3.0‐T scanners through in vivo studies. Numerous cardiac parametric mapping sequences were tested, including T 1 mapping with inversion and saturation preparation, T 2 mapping with 2D and 3D coverage, T 1 rho mapping, and simultaneous T 1 /T 2 mapping. The in‐line map‐building ability was further validated on an additional cohort of healthy volunteers, and the performance of map building was compared with the nonlinear least‐squares (NLS) curve‐fitting method. With BeatMapSeq, all testing sequences were interactively constructed on the scanner without programming and successfully applied to all volunteers. BeatMapSeq's MOLLI implementation showed excellent agreement with the corresponding vendor‐provided product. DeepFittingNet generated the corresponding maps fully automatically, without any user intervention, and the in‐line map‐building capability of BeatMapSeq enabled the parametric maps to be available on the scanner shortly after imaging. DeepFittingNet demonstrated good agreement with the NLS curve‐fitting method for T 1 , T 2 , and T 1 rho calculations. The in vivo T 1 , T 2 , and T 1 rho measured at 3.0 T in this study were consistent with the previously reported values obtained using the corresponding sequences. The proposed BeatMapSeq streamlined the sequence construction and map building for cardiac parametric mapping in cardiovascular magnetic resonance.