Premium
Isotropic reconstruction of a 4‐D MRI thoracic sequence using super‐resolution
Author(s) -
Reeth Eric,
Tan Cher Heng,
Tham Ivan WK,
Poh Chueh Loo
Publication year - 2015
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.25157
Subject(s) - isotropy , real time mri , computer science , magnetic resonance imaging , imaging phantom , image resolution , segmentation , sequence (biology) , visualization , artificial intelligence , computer vision , radiology , physics , medicine , optics , chemistry , biochemistry
Purpose Four‐dimensional (4D) thoracic magnetic resonance imaging (MRI) sequences have been shown to successfully monitor both tumor and lungs anatomy. However, a high temporal resolution is required to avoid motion artifacts, which leads to volumes with poor spatial resolution. This article proposes to reconstruct an isotropic 4D MRI thoracic sequence with minimum modifications to the acquisition protocols. This could be an important step toward the use of 4D MRI for thoracic radiotherapy applications. Methods In a postacquisition step, three orthogonal 4D anisotropic acquisitions are combined using super‐resolution to reconstruct a series of isotropic volumes. A new phantom that simulates lung tumor motion is developed to evaluate the performance of the algorithm. The proposed framework is also applied to real data of a lung cancer patient. Results Subjective and objective evaluations show clear resolution enhancement and partial volume effect diminution. The isotropic reconstruction of patient data significantly improves both the visualization and segmentation of thoracic structures. Conclusions The results presented here are encouraging and suggest that super‐resolution can be regarded as an efficient method to improve the resolution of 4D MRI sequences. It produces an isotropic 4D sequence that would be impossible to acquire in practice. Further investigations will be required to evaluate its reproducibility in various clinical applications. Magn Reson Med 73:784–793, 2015. © 2014 Wiley Periodicals, Inc.