Patient motion is a common challenge in the clinical setting and fast spin echo longitudinal relaxation time fluid attenuating inversion recovery imaging method with motion correction would be highly desirable. The motion correction provided by transverse relaxation time- and diffusion-weighted periodically rotated overlapping parallel lines with enhanced reconstruction methods has seen significant clinical adoption. However, periodically rotated overlapping parallel lines with enhanced reconstruction with fast spin echo longitudinal relaxation time fluid attenuating inversion recovery-weighting has proved challenging since motion correction requires wide blades that are difficult to acquire while also maintaining short echo train lengths that are optimal for longitudinal relaxation time fluid attenuating inversion recovery-weighting. Parallel imaging provides an opportunity to increase the effective blade width for a given echo train lengths. Coil-by-coil data-driven autocalibrated parallel imaging methods provide greater robustness in the event of motion compared to techniques relying on accurate coil sensitivity maps. However, conventional internally calibrated data-driven parallel imaging methods limit the effective acceleration possible for each blade. We present a method to share a single calibration dataset over all imaging blades on a slice by slice basis using the APPEAR non-cartesian parallel imaging method providing an effective blade width increase of 2.45×, enabling robust motion correction. Results comparing the proposed technique to conventional cartesian and periodically rotated overlapping parallel lines with enhanced reconstruction methods demonstrate a significant improvement during subject motion and maintaining high image quality when no motion is present in normal and clinical volunteers.
Copyright © 2012 Wiley Periodicals, Inc.