Premium
Accelerated MRI of the fetal heart using compressed sensing and metric optimized gating
Author(s) -
Roy Christopher W.,
Seed Mike,
Macgowan Christopher K.
Publication year - 2017
Publication title -
magnetic resonance in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.696
H-Index - 225
eISSN - 1522-2594
pISSN - 0740-3194
DOI - 10.1002/mrm.26290
Subject(s) - undersampling , compressed sensing , artifact (error) , computer science , magnetic resonance imaging , metric (unit) , artificial intelligence , iterative reconstruction , computer vision , pattern recognition (psychology) , medicine , radiology , operations management , economics
Purpose To develop and validate a method for accelerated time‐resolved imaging of the fetal heart using a combination of compressed sensing (CS) and metric optimized gating (MOG). Theory and Methods Joint optimization of CS and MOG reconstructions was used to suppress competing artifact from random undersampling and ungated cardiac motion. Retrospectively and prospectively undersampled adult and fetal data were used to validate the proposed reconstruction algorithm qualitatively based on visual assessment, and quantitatively based on reconstruction error, blur, and MOG timing error. Results Excellent agreement was observed between the fully sampled and retrospectively undersampled reconstructions, up to an undersampling factor of four. Visually, differences between ECG and MOG reconstructions of adult data were negligible. This was consistent with quantitative comparisons of reconstruction error (RMSE ECG = 0.07–0.13; RMSE MOG = 0.08–0.13), and image blur (B ECG = 1.03–1.20; B MOG = 1.03–1.20). The calculated MOG timing error (2–42 ms) was comparable to the acquired temporal resolution (∼60 ms). Quantitative evaluation of retrospectively undersampled (R = 2–8) fetal data (RMSE MOG = 0.06–0.12; B MOG = 1.04–1.27) was comparable to the adult volunteer results. Conclusion CS‐MOG for dynamic imaging of the fetal heart was developed and validated. Using CS‐MOG, images were obtained up to four times faster than conventional acquisitions. Magn Reson Med 77:2125–2135, 2017. © 2016 International Society for Magnetic Resonance in Medicine