Accelerated Breath-Hold Multi-Echo FSE Pulse Sequence Using Compressed Sensing and Parallel Imaging for T2 Measurement in the Heart

Introduction: Measurement of proton transverse relaxation time (T2) can be used to detect pathological changes in tissue for a variety of clinical applications, including identification of edema and iron overload. The most widely used T2 mapping pulse sequence is the Carr-Purcell-Meiboom-Gill (CPMG) sequence [1,2]. For body MRI, however, a CPMG pulse sequence is typically performed with navigator gating due to its low data acquisition efficiency. A new T2 mapping pulse sequence [3] based on multi-echo fast spin-echo (ME-FSE) readout with a net acceleration rate (R) of 3.2 (parallel imaging [PI] with R = 1.6 + turbo factor = 2) has been proposed to perform breath-hold (BH) body MRI acquisition. This new T2 mapping pulse sequence uses a reverse centric k-space reordering to sample the central half with more accurate even echoes and the outer half of k-space with less accurate odd echoes, and thereby provides even-echo CPMG accuracy. The new T2 mapping pulse sequence has been reported to yield spatial resolution of 2.7 mm x 3.8 mm within a breath-hold duration on the order of 20s. While this spatial resolution may be adequate for liver imaging, it may be marginally acceptable for cardiac imaging where myocardial wall thickness is typically on the order of 10 mm. We propose to further accelerate this pulse sequence using a recently developed joint reconstruction algorithm that combines compressed sensing (CS)[4] and PI to exploit joint sparsity of randomly undersampled data acquired from different receiver coils, as previously described [5]. Cardiac ME-FSE imaging is a good candidate for this joint CS-PI approach, since spatial and temporal correlations in the image series result in sparse representations. The purposes of this study were to highly accelerate the T2 mapping pulse sequence using the joint CS-PI technology and evaluate accuracy of the resulting T2 measurements.