Premium
TH‐EF‐BRA‐06: A Novel Retrospective 3D K‐Space Sorting 4D‐MRI Technique Using a Radial K‐Space Acquisition MRI Sequence
Author(s) -
Liu Y,
Subashi E,
Yin F,
Cai J
Publication year - 2016
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4958263
Subject(s) - imaging phantom , image quality , computer vision , computer science , k space , iterative reconstruction , artificial intelligence , frame rate , trajectory , nuclear medicine , real time mri , data acquisition , medical imaging , magnetic resonance imaging , physics , medicine , radiology , image (mathematics) , astronomy , operating system
Purpose: Current retrospective 4D‐MRI provides superior tumor‐to‐tissue contrast and accurate respiratory motion information for radiotherapy motion management. The developed 4D‐MRI techniques based on 2D‐MRI image sorting require a high frame‐rate of the MR sequences. However, several MRI sequences provide excellent image quality but have low frame‐rate. This study aims at developing a novel retrospective 3D k‐space sorting 4D‐MRI technique using radial k‐space acquisition MRI sequences to improve 4D‐MRI image quality and temporal‐resolution for imaging irregular organ/tumor respiratory motion. Methods: The method is based on a RF‐spoiled, steady‐state, gradient‐recalled sequence with minimal echo time. A 3D radial k‐space data acquisition trajectory was used for sampling the datasets. Each radial spoke readout data line starts from the 3D center of Field‐of‐View. Respiratory signal can be extracted from the k‐space center data point of each spoke. The spoke data was sorted based on its self‐synchronized respiratory signal using phase sorting. Subsequently, 3D reconstruction was conducted to generate the time‐resolved 4D‐MRI images. As a feasibility study, this technique was implemented on a digital human phantom XCAT. The respiratory motion was controlled by an irregular motion profile. To validate using k‐space center data as a respiratory surrogate, we compared it with the XCAT input controlling breathing profile. Tumor motion trajectories measured on reconstructed 4D‐MRI were compared to the average input trajectory. The mean absolute amplitude difference (D) was calculated. Results: The signal extracted from k‐space center data matches well with the input controlling respiratory profile of XCAT. The relative amplitude error was 8.6% and the relative phase error was 3.5%. XCAT 4D‐MRI demonstrated a clear motion pattern with little serrated artifacts. D of tumor trajectories was 0.21mm, 0.23mm and 0.23mm in SI, AP and ML directions, respectively. Conclusion: A novel retrospective 3D k‐space sorting 4D‐MRI technique has been developed and evaluated on human digital phantom. NIH (1R21CA165384‐01A1)