Reduced Blurring in 3D Fast Spin Echo through Joint Temporal ESPIRiT Reconstruction

Introduction: Volumetric (3D) Fast Spin Echo (FSE) is an attractive alternative to 2D FSE as it provides isotropic resolution. However, long echo trains are required to maintain scan efficiency, leading to blurring due to T2 decay [1]. Flip-angle modulation can reduce this effect [2], but blurring often persists, particularly in musculoskeletal applications. In this work, we aim to reduce blurring by modeling the signal’s temporal behavior. By constraining the temporal decay to a low-dimensional subspace as in [3] and by randomizing the echo train ordering as in [4], we are able to reduce blurring without explicitly estimating a T2-relaxation map. This approach combines ESPIRiT [5], a parallel imaging (PI) and compressed sensing (CS) method, with joint-thresholding of the temporal image coefficients to reconstruct a full time series of images. This differs from other approaches where CS and PI are combined in a parametric framework with the end goal of tissue quantification [3, 4, 6-9].